The problem with this automated approach is that it can be hard to tell when a collector has loaded the new remote configuration or if it experienced a configuration error.

Fleet Management solves this problem with the new effective configuration feature, which shows the remote configuration currently running on the collector. When you roll out a configuration change, the Configuration tab in the collector details drawer shows the rollout status. You can see both the server configuration and effective configuration while the rollout is pending. Once the changes successfully apply, only the effective configuration is shown. If an error occurs, the application flags the error and provides context for troubleshooting.

When you add a panel to a dashboard, you now have three starting points: an empty panel, a library panel, or a saved query. Select Use saved query to open the Saved queries dialog box and choose a query. Grafana then generates a ready-to-edit panel—including a recommended visualization—so you can add a useful panel to your dashboard without even entering panel edit mode:

To try out this feature, enable the queryLibrary feature toggle.

The Saved queries drawer has been replaced with a dialog box, providing more space for query titles, data source names, and other query details. The individual query cards also have a refreshed design, allowing for easier readability.

The Favorites tab has been replaced with a Starred queries filter view and any search, filter, or sort options you set on the All view apply to this view as well.

We’ve added a Remember filters switch, which lets you persist your filter selections across sessions in your local storage. With this addition, you no longer have to reset filters or searches that you consistently use every time you come back to saved queries.

Finally, the lock/unlock option and the Recent tab have been removed.

These updates deliver a cleaner interface and a more consistent way to discover, manage, and reuse saved queries.

With Database Observability, you can identify slow and failing queries, understand wait events, analyze explain plans, and examine query samples all within Grafana. Connect your databases through Grafana Alloy and get instant visibility into what’s happening across your fleet.

Database Observability supports MySQL and PostgreSQL across all major cloud providers and self-managed environments.

You can now see how much each stack contributes to your organization’s monthly bill, alongside the existing usage data

Key features include:

• Cost visibility by stack: Understand how each stack contributes to overall spend
• Expanded cost data: New columns for Amount Due, Included Usage, Overage, Attributed Usage, and Attributed Cost
• Organization vs. stack breakdown: Clear separation between organization-wide totals and per-stack contribution
• Improved accuracy for user-based products: Attributed Usage ensures costs are calculated using deduplicated users across stacks

Coming soon:

• In addition to the organization wide CSV export, you’ll soon be able to explore usage and cost for a specific stack directly within the Cost Management and Billing app, making it easier to analyze spend

How to access:

The updated export is available via the API (GET /api/orgs/:slug/billed-usage?format=csv) for users with the org-billing-info:read permission, or in the account portal under Billing → Billed usage history

You are affected if

You use the GET /api/alertmanager/grafana/api/v2/status endpoint and rely on the alert.notifications:read permission to access it.

Migration

If you have custom roles that need access to this endpoint, add the alert.notifications.system-status:read action to those roles. Admin users are unaffected as they receive this permission automatically through the built-in alerting notifications writer role.

To try out panel styles, enable the vizPresets feature flag.

• Health page shows a live snapshot of active issues across all your Clusters. The Risks tab surfaces reliability and availability problems across twelve checks, while the Efficiency tab highlights resource configuration gaps such as missing requests or limits and over-provisioned CPU or memory.

• Flame graphs are embedded in Workload and Pod detail pages. You can investigate which parts of your code are consuming the most CPU or memory without leaving Kubernetes Monitoring.

• Simplified configuration page showing four basic steps to configure using the Kubernetes Helm chart.

We’re excited to announce the general availability of Attribution Alerts in the Cost Management and Billing app!

Key features include:

• Label scoped alerts: Create alerts filtered by attribution labels like team, service, or environment
• Flexible thresholds: Define alert conditions based on usage metrics such as log volume (GB) or active series
• Multi label filtering: Scope alerts using one or two attribution labels for more precise targeting
• Integrated alerting: Send notifications through existing Grafana Alerting contact points
• RBAC support: Ensure proper access control for creating and managing alerts

To learn more, check out our documentation.

The Alert activity menu item now contains two tabs, one for Alerts, and another for Active notifications.

The Alert rules submenu takes you to the main Alert rule page, which now includes the Recently deleted rules in their own separate tab on the page.

The new Notification Configuration submenu item takes you to a page where all of your notification-related content is available. This includes tabs for Contact points, Notification policies, Time intervals, and Templates all in one easy to access page.

Annotation clustering

Group adjacent point annotations into a single combined region with the Annotation clustering switch. Clustered annotations aggregate their contents into a single, scrollable tooltip, so you can explore dense annotation data without overwhelming the panel.

To try this feature, enable the annotationsClustering feature flag.

Annotation indicator controls

Control how annotations appear in your panels with the Hide lines and areas switch. With this feature, you can hide annotation indicator lines and shaded regions to reduce crowding in panels with a high number of annotations.

To try this feature, you must have the annotationsClustering feature flag enabled.

Multi-row annotations

Display each annotation query on its own row within a panel, so you can clearly distinguish between annotation sources with the Multi-row annotations switch. The multi-row option appears automatically when your dashboard includes more than one annotation query.

With this update, you can pin annotation tooltips with a click for easier interaction. We’ve also made annotation markers keyboard-accessible, so you can tab through many annotations without a mouse.

This feature is generally available.

Cloud Provider Observability now lets you bring your own dashboards, offers you the ability to customize instance drilldown view panels, and make these assets available across all your teams. Everyone can monitor cloud services with the same trusted views, while still having the flexibility to tailor what they see.

Make dashboards your own

Attach custom dashboards to any cloud service and set them as the default. Your teams land on the right insights instantly with no extra clicks, or hunting for context.

Tune instance views to your needs

Customize instance-level drilldowns for supported services like RDS and Azure Virtual Machines. Choose the metrics you care about and configure units so data is always clear and actionable.

Jump straight into deeper analysis

A new shortcut takes you directly to Metrics Explore in Grafana Cloud, making it easy to go from high-level monitoring to detailed investigation in seconds.

Access new configure page from KPI drawer

When you select an entity in the Knowledge graph, the KPI drawer now includes a direct link to that service’s configure page in Cloud Provider Observability. Use the drawer to jump from a metric or alert to the full service view — including your custom dashboard and instance drilldown — so you can go from spotting an issue to understanding its context in one click.

For setup steps and detailed guidance, see Dashboard and views settings in the Cloud Provider Observability documentation.

Grafana Assistant in Slack allows you to access the data and knowledge in Grafana Assistant in Slack. Ask about your traces, metrics, the latest deployment and how to fix high p99 latency in your search service.

Invite your colleagues to collaborate with you: They can tag @Grafana in your thread and engage with you and the Assistant in the thread. This allows to bring together knowledge across teams and squads easily while maintaining one conversation.

Grafana Assistant in Slack becomes even more powerful when you hook up more data sources in Grafana Cloud. This allows you to correlate telemetry with business data in Slack.

• How much did us this incident cost?
• Does churn increase when loading times are higher than 5s?
• Based on historical records, how much load can we expect next Monday?

Watch this video to get a feel how Grafana Assistant works in Slack.

Skills let you write runbooks, workflows, hints and context guides so that the Assistant can drill down even better in your stack. You can also associate slash commands with Skills so that /my-skill brings up your Skill and the Assistant works through it specifically.

Within Skills, you can even reference your integrations like GitHub, Jenkins, Cloudflare and many more to have Grafana Assistant act across your whole DevOps and SRE stack.

The Assistant continuously discovers Skills on its own so you can reference them in your questions but you don’t need to necessarily. When talking about your payment service, the Assistant will retrieve the appropriate runbook and its instructions for you and run through them.

To address this, Grafana now includes a Limit option in the legend settings. Set a numeric value to control how many series items appear in the legend by default. When the series count exceeds the limit, a Show all link allows you to see the full list of series.

The visualization itself is unaffected—all series continue to render in the chart regardless of the legend limit.

The Limit option is available in the legend settings for the following visualizations: time series, bar chart, bar gauge, candlestick, histogram, pie chart, state timeline, status history, trend, and XY chart.

Use Grafana Assistant right in the CLI alongside all the other tools you use for your observability job. In the CLI, the Assistant can help you with all the things it can also do in the browser or in Slack. Ask about recent errors, latency or traces - get the answer in the CLI. With the CLI, Grafana Assistant can also read, create, update and delete local files for you: bring observability, coding and configuration closer than ever in one tool without the need to context switch.

Additionally, Assistant Tunnel, a part of the CLI, enters Public Preview too. The Tunnel functionality allows you to link up your Grafana Assistant web experience with your local machine. Now the Assistant can run kubectl, systemctl or any other CLI tool in Grafana Cloud for you. Look at error logs and latency panels while you simultaneously run CLI commands to have the Assistant diagnose root causes faster than ever.

Get the Grafana Assistant CLI on GitHub (full usage requires an active Grafana Assistant subscription): https://github.com/grafana/assistant-cli

If your workflow depends on inspecting the TO header of alert emails, contact Grafana Support to have BCC disabled for your instance.

From a single view, you can see all stored secrets, assign decrypters so only approved services can use each secret, and control access with granular RBAC permissions. This helps you reduce hardcoded credentials, rotate secrets with less operational overhead, and enforce least-privilege access as usage scales.

The centralized Secrets Management UI is currently available in Public Preview, access it today in your Grafana Cloud instance at Administration > Secrets Management.

Previously, if a saved query contained variables that weren’t in your dashboard, Grafana displayed a compatibility warning and limited how the query could be reused. With this update, the warning is replaced with a section where you can update the variable (for example, ${job}) inline by mapping the variable to either:

• A variable in your dashboard
• A custom value that you enter

Grafana applies your selections to the query before inserting it into the dashboard. However, the substitutions only apply to the query when it’s reused, and the original saved query remains unchanged.

This makes saved queries more portable and reusable in environments where variable names or available values differ between dashboards.

This functionality is also available in Explore for custom values.

For more information, check out the video demo:

Back in August of last year, we completed work to seamlessly migrate the code behind the Grafana Cloud homepage to a new plugin and internal code repository, as part of a larger effort to consolidate related code. Following this migration, we updated the code in the legacy plugin to ensure that accessing the old homepage at /a/cloud-home-app within a Grafana Cloud stack would automatically redirect you to the new homepage at /a/grafana-setupguide-app/home.

We’re now moving ahead with the final step in cleaning up this legacy plugin by removing the redirect and the plugin itself from all Grafana Cloud stacks. Accessing the old homepage at /a/cloud-home-app now results in an App not found error message with a link to the new homepage. If you already access your Grafana Cloud homepage by going to the root URL (e.g., <YOUR-STACK>.grafana.net), no action is needed as this will continue to serve the new homepage. If you’ve still got links saved that include /a/cloud-home-app, update those to go to the root URL instead.

Templates are a powerful way to create dashboards quickly, but they don’t always match your real-world metric names or data sources. The Assistant bridges that gap by:

• Identifying what the template expects (panel types, description, likely data).
• Matching those expectations to the data sources you actually have.
• Creating a dashboard tailored for your environment (and guiding you through any follow-up choices).

This reduces work and makes it faster to create a working, relevant dashboard from a template without the need to manually map metrics and panels.

After announcing the Public preview for Grafana Cloud at the beginning of the month, we are now enabling more functionalities for Git Sync. The top voted feature request was enabling the connection with more providers, and also with self-hosted instances of Git repositories. Today, we are bringing you 3 additional connection types to Grafana Cloud: GitLab, BitBucket and a pure Git integration that adds universal provider support to Git Sync.

The pure Git features is also available in Grafana OSS, and the other two providers are available in Grafana Enterprise.

To learn more, please consult our documentation or watch our latest YouTube video.

As usual, the feature is rolling out to all Cloud users and will become available as your stack’s release channel receives the update.

Labels are shared across Alerting, IRM, and SLOs. A label you create once, like service_name:payments, works everywhere: routing alerts, categorizing incidents, and correlating SLOs with the services they monitor. Service Center uses those same labels to give you a unified view of service health, team ownership, and related services.

There’s nothing to migrate. Your existing labels are already available in the Label management UI.

To get started, refer to Manage labels for alerts, incidents, and SLOs.

Previously, all environment variables from the Grafana process were automatically forwarded to plugin processes. Now, plugins only receive:

• Grafana-specific environment variables, such as GF_VERSION, GF_EDITION, GF_APP_URL.
• A limited set of permitted host environment variables used for HTTP proxy configuration, such as HTTP_PROXY, http_proxy, HTTPS_PROXY, https_proxy, NO_PROXY, and no_proxy.
• Plugin-specific configuration environment variables, such as AWS settings, Azure settings, and tracing configuration.

Migration

If your plugins require access to additional host environment variables, configure Grafana to forward all host environment variables to specific plugins using the forward_host_env_vars configuration option.

Add the following to your grafana.ini file under the [plugins] section:

ini Copy

[plugins]
forward_host_env_vars = plugin-id-1,plugin-id-2,plugin-id-3

Only add plugins to forward_host_env_vars if they require access to host environment variables. Forwarding all environment variables reduces the security isolation between Grafana and plugin processes.

When Grafana v12.4 starts, the migration system automatically transfers your playlists from the legacy SQL database to unified storage and validates data integrity after the migration completes.

We’ve built this system to run independently for each organization in your Grafana instance, ensuring reliable and isolated migrations. Unified storage brings playlists in self-managed Grafana to parity with Grafana Cloud, providing a modern, consistent storage backend for your resources. This is the first step in a broader effort to migrate all Grafana resources to unified storage—folders, dashboards, and library panels will follow. After playlist migrations are complete, the legacy playlist table will be deprecated.

This automatic migration is part of our ongoing work to enable the experimental Git Sync feature. After your instance is migrated, you can try out Git Sync by enabling the provisioning feature toggle.

How does this work?

At startup, Grafana checks the number of folders and dashboards in your instance. If the counts fall below a configurable threshold, Grafana automatically migrates these resources and switches them to the new storage mode. By default the threshold is set to 10 for both dashboards and folders.

Larger instances continue to operate as before, with no automatic changes applied.

What to expect

• Small instances with up to 10 dashboards or 10 folders are automatically migrated on startup.
• Larger instances are not affected. You can opt in manually when you’re ready by setting the following in your grafana.ini or custom.ini configuration file:
  ini Copy

  [unified_storage.folders.folder.grafana.app]
  enableMigration = true
  
  [unified_storage.dashboards.dashboard.grafana.app]
  enableMigration = true

• If any migration issues occur, Grafana logs a warning so you can investigate early. Full enforcement is scheduled for a future release.

Seegmentation to granularly manage Adaptive Logs is now Generally Available. Viewing recommendations, bulk editing, and understand savings impact are now all streamlined into a single-segment specific view. SRE’s can feel more comfortable optimizing safely within team boundaries, and balancing the optimization load with their application teams.

In Grafana v13, we’ve removed the duplicated metrics grafana_caching_items and grafana_caching_size.

Mitigation

We recommend you update your dashboards to use the following metrics instead:

• grafana_caching_resource_cache_size
• grafana_caching_query_cache_size
• grafana_caching_resource_cache_items
• grafana_caching_query_cache_items

The assistant uses details from sessions with the selected error to identify user behavior leading up to errors and patterns in error logs, giving you actionable insights to resolve issues faster.

This feature is generally available now, with more automatic root cause analysis and error resolution capabilities coming soon.

Saved queries are now shared by default: all users with access can reuse them, while only users with the Writer role can create, edit, or delete them. This replaces the previous per-query sharing checkbox with consistent, role-based permissions.

Two custom roles are available:

• Writer: Create, update, and delete saved queries
• Reader: Reuse existing saved queries

This makes it easier to collaborate on saved queries without risking accidental changes, while keeping ownership and control clear.

This new capability allows administrators to define which fields (or columns) are displayed by default based on the selected set of labels. Instead of relying on a static column setup, the logs view can now support a pre-configured default state per service, or set of services, ensuring users immediately see the most relevant information for their use case.

With this change, we aim for an improved log exploration with focus on the most important fields, providing a more consistent and guided experience across teams.

This enhancement helps reduce manual column adjustments and ensures that log drilldowns start with the right context every time.

Saved explorations include your active filters, so you don’t have to recreate the same view each time you investigate an issue.

How this works depends on your deployment:

• Grafana OSS: Explorations are saved in your browser’s local storage. They’re available on the same browser and device, and are cleared if you clear browser data.
• Grafana Cloud and Grafana Enterprise: Explorations are stored as saved queries. They persist across sessions and browsers and are tied to your user account, and can be optionally shared across the organization.

This makes it easier to revisit common investigations, keep ongoing debugging sessions, or maintain a set of frequently used log views.

Diversity sampling captures at least one trace for each unique trace combination (fingerprint) within a fixed time window. These combinations are defined by attributes such as service, route, and status code. With this baseline coverage in place, you can reduce probabilistic sampling rates and avoid creating and managing additional policies just to make sure developers find the traces they need.

You can restore dashboards you’ve deleted, no support ticket needed. With the new Recently deleted view, accidentally removing a dashboard no longer means lost work or rebuilding panels from scratch.

When you delete a dashboard, it now moves into a deletion history area instead of being removed permanently. From there, you can browse previously deleted dashboards and restore them to any folder you choose.

Restore dashboards from the new Recently deleted view

From the Dashboards page, click Recently deleted to see dashboards you’ve deleted. Admins can view all dashboards removed in the last 12 months. 

Select one or more dashboards to restore. The folder behavior is as follows:

• If a dashboard’s original folder still exists, it’s preselected.
• If multiple dashboards were deleted from the same folder, that folder is preselected.
• If they came from different folders—or the original folder no longer exists—simply choose a new target folder.

Restored dashboards inherit the permissions of the folder you restore them into, making it easy to get them back into the right place.

How deletion history works

Deleted dashboards are stored for up to 12 months, with space for 1,000 dashboards per org. When that limit is reached, the oldest entries are automatically removed, even if the full retention period hasn’t passed.

Availability and limitations

This feature is currently available to Grafana Cloud only, and will roll out to customers starting February 2026. For more information, refer to the documentation. 

As mentioned above, we would love your feedback, you can share it here.

Current limitations:

• Dashboard-level permissions are not restored; restored dashboards inherit folder permissions
• Version history is not preserved (restored dashboards begin at version 1).
• Folders themselves cannot be restored.

• Spending across the past 60 days
• Potential savings for the next 30 days
• How much you can save by optimizing resources
• Recommendations from Grafana Assistant about your data

This feature, now starting to roll out to all Grafana Cloud instances in Public Preview, lets you benefit from a native Git workflow directly from Grafana.

With Git Sync, you can combine the simplicity of editing dashboards in the Grafana UI with the knowledge that your dashboards are stored in a remote repository, and changes are traceable. Try it now in your Grafana Cloud instance.

To learn more about Git Sync, read this blog post or watch this video.

We’ve improved that experience.

You can now set a default spreadsheet when creating a Google Sheets data source, so it’s automatically available whenever the data source is used. Instead of repeatedly selecting the same file, users can define it once and reuse it across dashboards, panels, and queries.

The result is a smoother workflow for common use cases like tracking KPIs in shared spreadsheets, visualizing operational checklists, or building internal dashboards from collaboratively maintained data. Teams that rely on a single spreadsheet as a source of truth can now treat it as a stable data source instead of a repeated selection step.

This feature is supported for the Custom and Query variable types. For query variables, you can create multi-property variables for Infinity and PostgreSQL data sources.

This feature is being rolled out across Grafana Cloud over the next few weeks. To try it out in self-managed Grafana, enable the multiPropsVariables feature toggle.

How it works

Define variable options as a JSON array of objects. Each object can have any number of properties. For example, you might want to visualize data across multiple environments using properties such as text, value, aws, azure, or google. You would configure this data as follows in the custom variable configuration:

[ Copy

     { "value": "1", "text": "Development", "aws": "dev", "azure": "development", "google": "googledev" },
     { "value": "2", "text": "Staging", "aws": "stag", "azure": "staging", "google": "googlestag" },
     { "value": "3", "text": "Production", "aws": "prod", "azure": "production", "google": "googleprod" }
]

This is how those values are displayed in a preview:

Then use ${var.someProperty} to reference any property in your dashboard panels:

You can even deeply nest properties and still access them using familiar variable syntax. In the following example, each user has an address property with all the elements of an address nested within it:

User interface updates

To support this feature, we’ve updated the user interface.

We’ve added a values format selector to the custom variables editor so you can choose between CSV and JSON, and author JSON object options directly.

For the custom and query variable types, the editor now also shows a Preview of values table, including the properties, so you can quickly verify property names and values. The preview displays deeply nested properties, and the editor supports autocomplete for property paths.

If you are using more than one integration in you observability fleet this new functionality will help you understand and manage their state.

It will enable you to:

• Update them with a single click.
• Navigate to their dashboards and alerts.
• Find their documentation.

The three summary cards on top that give you an instant overview of your integrations health. They show you how many of:

• Integrations you have.
• Are running the latest version.
• Need an update.

The new table view helps you wave off complexity to view your complete Integrations estate. Each row shows:

Everything you need is visible in one row. It’s now easier to find the dashboards and alerts you need without tab switching.

One-click updates

When a new version of an integration is available, you’ll see an “Update” button. You can click it, confirm in the modal, and you’re done. Then you’ll see “Up to date” with a green checkmark in the same place confirming that the update is successful.

Try it out

The new Manage Integrations page is available now.

Head to Connections > Manage integrations in your Grafana Cloud instance to check it out.

If you have feedback, drop by the #integrations channel in the Grafana Community Slack.

We’d love to hear from you!

You can now automate your entire user lifecycle and team structures directly from your identity provider (IdP). By moving SCIM from public preview to general availability, we’ve removed the “first login” bottleneck, allowing centralized operators to provision, configure, and secure users before they even open the app.

Instead of managing complex sync pipelines and waiting for users to log in to assign permissions, you can now build, test, and scale your entire engineering organizational structure in minutes.

• Onboard instantly: Users and teams are created the moment they are added to your IdP, enabling immediate configuration in tools like IRM.
• Automate deprovisioning: Gain peace of mind with automatic seat reclamation and instant access removal the second a user is deactivated in your IdP.
• Scale faster: Early adopters of SCIM see more active users within 90 days compared to SAML-only customers, because removing manual friction accelerates platform adoption.

Enable SCIM now - available with Okta and Entra ID for Grafana Cloud and Enterprise v12.4.

Learn how to configure SCIM

Secrets management gives you a centralized place to securely store sensitive data like API keys, passwords, and tokens.

You can reference secrets by name in your scripted and browser Synthetics checks so that sensitive values are never visible or stored as cleartext. Secret values are also automatically masked in all Synthetic Monitoring output to prevent accidental exposure.

Currently secrets can only be accessed in scripted and scripted browser checks, support for single HTTP will be available in the future as we continue to iterate on the feature.

To learn more, refer to the secrets management documentation.

• Fleet composition – Visualize your collector types, versions, and operating systems at scale
• Resource usage – Track your current collector and pipeline counts against stack limits
• Troubleshoot at a glance – Spot pipeline errors across your fleet instantly
• Active deployments – Monitor total collectors and active pipelines in real-time

No matter the size of your fleet, the Overview tab puts critical metrics front and center, so you can identify trends, spot issues, and make informed decisions faster than ever.

These enrichments can perform actions like querying data sources (metrics and logs), initiating background tasks like Assistant Investigations, adding direct links to tools such as Sift and Asserts, or even using an LLM Plugin to explain the alert’s context. Alert enrichments are scoped by either labels, annotations, or applied across all alerts.

Check out our documentation and experiment with alert enrichment today and see it empower your workflow immediately!

Multi-device support changes that. You can now connect multiple phones to your Grafana IRM account and receive push notifications on all of them simultaneously. Switch between devices without missing a beat. Keep a backup phone connected when your primary is charging. Stay reachable wherever you are, on whatever device you have at hand.

Your on-call schedules, alert groups, and incidents sync seamlessly across all connected devices. Acknowledge an alert on one phone, and it updates everywhere. No more juggling logins or worrying about which device is “the active one.”

Get started by downloading the latest version of the Grafana IRM mobile app and signing in on any additional devices you’d like to connect.

When you preview a visualization using the updated Suggestions feature, we give you a full-sized preview of the suggestion in the panel editor before you select the suggestion.

The quality of suggestions provided should improve in the updated version of the feature, as we now leverage information provided by data sources, such as the data frame type found in the data plane, to make educated choices on which visualizations are the most suitable. We’ve also tuned the system overall and updated the user interface to make the suggestions clearer.

To try out the updated visualization suggestions experience, enable the newVizSuggestions feature toggle.

Geographic analysis is now built directly into Frontend Observability, making it easy to validate location-based performance hypotheses without custom queries or additional tools. With the Geolocation tab, geographic context is a first-class way to investigate frontend issues and validate location-based hypotheses.

Core Features:

• Visualize performance and error hotspots on an interactive, color-coded map
• Drill down from global to local regions to isolate problem areas
• Switch between Core Web Vitals and Error Rate to compare by different user experience KPIs
• Jump directly from a region into filtered error or performance views

Use Cases

• Find regions with slow loads or high error rates after a deployment
• Validate CDN placement and regional infrastructure decisions
• Detect localization or network-specific issues during location-based rollouts
• Prioritize performance optimizations based on real user impact by region

To use this Geolocation Map feature, you will need to turn on geolocation for your application. Learn how to get started in our documentation.

The Circular shape has been introduced as an alternative to the existing Arc gauge under the new Style option.

We’ve also added sparkline support to the gauge, which improves the its usefulness as an alternative to the stat visualization for time series data.

Gradient support was also added. Gradients don’t just look nice—they help users interpret panels with thresholds, or with series that use a continuous color scheme.

The new gauge also has accessibility improvements. The Center glow option renders the gauge’s current color value as a transparent glow behind the gauge’s text. This replaces the previous behavior of changing the color of the text itself. This change helps avoid situations where text can become illegible due to an individual user’s theme.

There are several more new configuration options, including:

• Rounded bars (under Bar style)
• Configurable bar thickness
• Endpoint markers for readability
• New text display modes to precisely configure the information presented in your visualization

To try out the revamped gauge visualization, enable the newGauge feature toggle.

We’ve added new ways to travel through time in dashboards and panels to make exploring your metrics and data more intuitive and efficient.

Use dashboard-level controls to jump one full time span backward or forward. For more precise time travel, we’ve now made the x-axis of time series visualizations interactive at the panel level as well. Just click and drag the x-axis timestamps horizontally to move forward or backward in time; then release to update the dashboard time range.

To zoom in on a time range, click and drag in the visualization area from a start time to any end time, and then release. To zoom out, double-click on the visualization area to expand the time range by half on either side of the center, showing twice as much time.

Finally, two new time range keyboard shortcuts have been added to Grafana. Typing t+ zooms in to half the current time range, and conversely t- zooms out to twice the current time range.

Time range pan and zoom applies to all visualizations that support time series data:

• Candlestick
• Heatmap
• State timeline
• Status history
• Time series

Grafana Assistant is the premier AI feature in Grafana Cloud that lets you create dashboards, query data, analyze alerts, manage incidents, improve instrumentation and much more. With the Assistant, you can ramp up faster than ever.

Grafana Assistant in Slack is the same experience but in Slack so that you can chat with Grafana Assistant anywhere and at any time. Collaborate with your colleagues on data and incidents, get quick updates about anything without leaving your team channel.

Start your first conversation by using @Grafana after installing the Slack app.

Every user needs an active Grafana Assistant user to connect to the Slack app.

Share Dashboards with Anyone

You can now share your Grafana Zabbix dashboards externally without requiring recipients to have access to your Grafana organization. This breaks down collaboration barriers, making it easier to share critical monitoring data with stakeholders, clients, or external teams.

Enhanced Stability and Performance

We’ve introduced improvements to ensure more reliable and efficient querying. New query guardrails provide proactive warnings when a query might exceed Zabbix API limits, helping you prevent performance degradation and unexpected resource usage. You have now more control by explicitly setting query timeouts in the configuration, preventing long-running queries from impacting your system.

Granular Filtering with Host Tags

A highly request new feature is Host Tags for granular filtering. You can now use Zabbix host tags directly in your Grafana Zabbix queries, enabling more granular, flexible data filtering and visualization than ever before, allowing you to focus on the exact data you need, faster.

New functionality

Tabs

Structure your dashboards horizontally, as well as vertically, to organise the content for different contexts and user groups.

Show / hide functionality

Show or hide panels, rows, or tabs based on the selected variable or whether the panel returns data, so viewers only see what’s relevant to their context.

Auto grid layout

Set your row, tab, or dashboard to Auto grid so the panel size adapts to the space available in the display. Configure minimum column width, row height, and the maximum number of columns per row to meet your requirements. If you prefer setting different panel widths and heights within the same row, you can use the Custom layout option, which is the same as the pre-Dynamic dashboards grid. Auto grid also enables the show/hide functionality. (To avoid leaving unsightly gaps in the dashboard, we require you to set the layout to auto grid before configuring any show/hide rules.)

Dashboard content outline

Find the content you need in your dashboards easily through a clickable outline. As an editor, click on an element in the outline to scroll to it and select it to make quick changes to the title or description. As a viewer, click on a panel, row, or tab to navigate to it instantly and find the information you’re looking for.

User experience changes

Side toolbar instead of a second top toolbar

To provide additional vertical space, we’ve removed the second top toolbar in favour of a toolbar on the right side of the screen. This helps both viewers and editors see more information at once. You can find the dashboard content outline and alert rules in the side toolbar, and editors can also access dashboard settings there.

Context-aware editing from the dashboard

Quickly update a panel’s title and description from the dashboard sidebar without entering panel edit mode. This sidebar also lets you quickly edit rows, tabs, and template variables.

What changed

OpenTelemetry log producers send less information in the log body and more data as structured metadata, or key–value pairs for resource attributes, log fields, and log attributes. Previously, much of this information was only visible when expanding a log line, making important context easy to miss.

Grafana now detects OpenTelemetry log lines and surfaces relevant metadata directly in the log view.

How OpenTelemetry log lines are identified

A log line is treated as an OpenTelemetry log when it contains OpenTelemetry-related fields, such as severity_number. For these logs, Grafana applies OpenTelemetry-specific rendering logic by default.

What is shown by default

For OpenTelemetry log lines, Grafana now displays the following fields inline when present:

• Thread name
• Exception type
• Exception message
• Log message (body)
• Log attributes (new)

Along with these fields, the logs panel displays by default the corresponding timestamp and a colored level, for every log line, OpenTelemetry or not.

Log attributes

The new Log attributes field groups together key–value pairs from the log metadata that are likely to be application-specific and relevant for debugging.

It includes metadata fields that do not match the following categories:

• OpenTelemetry resource attributes. For example, infrastructure and platform metadata such as k8s\_\*, cloud\_\*, host\_\*, service\_\*.
• OpenTelemetry internal log fields such as severity, trace or span IDs, timestamps, and flags.

This means infrastructure-level and internal OpenTelemetry fields remain available in log details, while application-level metadata is promoted to the main log line view.

Suggested fields in the field selector

To make these fields easier to work with, Grafana now surfaces suggested fields in the field selector shown to the left of each log line.

For OpenTelemetry logs, the suggested fields include when present:

• Thread name
• Exception type
• Exception message
• Log attributes
• Scope name

These suggestions help you quickly add, remove, or reorder the most relevant OpenTelemetry fields without searching through the full field list.

Why this matters

Less empty log lines: As log bodies become smaller, important context no longer disappears into collapsed metadata.

Better signal at a glance: Exception details and meaningful attributes are visible without expanding every log entry.

Cleaner defaults: High-volume, low-value metadata stays out of the way while still being accessible when needed.

Overall, this change makes OpenTelemetry logs easier to scan, faster to understand, and more useful by default, especially in high-cardinality or highly structured logging environments.

• Prevent Bill Shock: Automatically block expensive, inefficient queries before they execute.
• Optimize Performance: Keep your environment fast by ensuring large-scale searches don’t consume excessive resources.
• Actionable Feedback: Users receive clear error messages when a query is blocked, guiding how to refine their search (e.g., narrowing the time range).

This feature is still behind a feature flag, and reach out to support to enable it!

After you successfully configure a data source, click the Build a dashboard button:

The empty dashboard that opens automatically displays a list of suggested dashboards based on the data source:

This allows you to deploy a functional dashboard faster than building a brand new one.

Suggested dashboards provide:

• Contextual discovery: Automatically suggests pre-built dashboards tied specifically to the data source you have connected.
• Accelerated setup: You can immediately load and save suggested dashboards, which can then be used as a starting point for further customization.
• Reduced friction: Streamlines the onboarding process by offering actionable content directly after data source configuration.

• All incident updates are organized in one place.
• Team members can follow along on incident progress without leaving Slack.
• Stakeholders can subscribe to announcement threads using Slack’s Get notified of new replies setting.

Configure incident announcements to route incidents to the right Slack channels, then subscribe to announcement threads for the incidents you want to follow.

After it’s installed, Instrumentation Hub automatically discovers what’s running across your environment, building a catalog of services organized by namespace and technology so teams can immediately see what exists, what’s instrumented, and what isn’t.

You can also remove instrumentation.

Now you can avoid:

• Slow, brittle, manual rollouts
• Indiscriminate auto-instrumentation that over-collects, drives up cost, and obscures signal quality

Proven, out-of-the-box pipelines are deployed automatically, while decisions about what to instrument and how deeply remain explicit and reversible.

When you configure query variables with regex for data sources like Azure, CloudWatch, and AWS IoT, the regex matches the value. However, this value is often different from the display text, which makes it difficult to match the value you need.

To fix this mismatch, use the Apply regex to option in your query variable settings to select whether the regex applies to the value or the display text.

To learn more, refer to the variable documentation.

A red health status on the Inventory page tells you something’s wrong. When you click on the collector, you see the Configuration tab and a list of configuration errors, complete with row and column numbers. Identify the faulty pipeline, click on it in the Connected Pipelines list, and make your fix. With Fleet Management, you can spot configuration errors at a glance and fix them quickly, without ever leaving the application!

Upload your source maps directly to Grafana so that you do not need to host these publicly and risk exposing your IP. Get started by including the bundler plugins into your build pipeline. Learn more by going to your Frontend Observability App -> Settings -> Source Maps.

What’s new in this GA release:

• New documentation for troubleshooting issues with uploading Source Maps.
• Improved support for NextJS applications.
• Support uploads from Azure regions which were previously blocked.
• Uploads now include application version to make uploads easier to distinguish in the source maps list

This new menu consolidates variables, links, and annotation controls into a single collapsible drop-down list. When any of these items are configured to appear in the menu, Grafana shows a dedicated toolbar button with a count of how many controls are available. Clicking the button opens a neatly structured menu where variables appear first, followed by annotation controls and links, each grouped for easy navigation.

You can now choose to hide variables, annotation controls, or dashboard links from the main toolbar and place them in the dashboard controls menu instead.

SLOs can spike above 100% for a few reasons:

• SLO is not in a ratio format that provides a value between 0-1
• Ratio queries that use different metrics in the numerator and the denominator can experience synchronization issues when the numerator and denominator are incremented at different intervals
• Adaptive Telemetry can cause impact on recording rules, which can lead to unexpected irregularities in recorded SLI data
• The ruler executes recording rules with varying latency. SLI series may be inaccurate or report over 100% when the numerator and denominator of the ratio query come from different recording rules
• Counter series that transition from being absent to being present cause discrepancies in the SLI calculation when the series’ initial value is not zero

Users can now select to clamp their panels on individual SLO dashboards to prevent spikes over 100% from impacting their view. To help users identify and troubleshoot SLI query issues, an annotation is added to the SLI panel when spikes above 100% occur. The annotation points to documentation that guides users through analysis and troubleshooting steps to identify why the spike is occurring.

Press Cmd + I to open the capture flow, take a screenshot, and attach it to an existing incident. The screenshot, along with a source link, appears on the incident page and in the incident’s Slack channel so responders and stakeholders stay aligned.

For more information on this enhancement, see the preconfigured dashboards and alerts pages for AWS, Azure, and GCP in the Cloud Provider Observability documentation.

As a part of the change, the version shown in the Grafana Help menu and on the login page will now simply display “Grafana Cloud”—better reflecting that updates happen continuously across Grafana Cloud.

You don’t need to take any action. This change will roll out gradually over the coming months as we continue to make Grafana Cloud updates more reliable and timely.

Google Cloud Platform (GCP):

• Cloud Run: pinpoint latency or error spikes in Cloud Run apps.
• Cloud Storage (GCS): track storage access patterns and quotas in Cloud Storage (GCS).
• AlloyDB for PostgreSQL: monitor AlloyDB performance trends across compute and I/O.

Microsoft Azure:

• SQL Database - Elastic Pools: right-size your Elastic Pools with real-time and historical DTU usage.
• PostgreSQL Flexible Servers: Catch slow queries and IOPS bottlenecks in PostgreSQL workloads.

Amazon Web Services (AWS):

• Bedrock AgentCore: health tracking and slow operation detection for Bedrock AgentCore.
• DynamoDB: hot key detection for DynamoDB.
• ElastiCache: memory usage trends for ElastiCache.
• Route 53: DNS health checks and resolution time breakdowns for Route 53.
• Route 53 Resolver: DNS health checks and resolution time breakdowns for Route 53 Resolver.

These dashboards are available to install with one click from the Cloud Provider Observability integration pages for each cloud. Jumpstart your monitoring strategy and start detecting issues before they impact your users.

For full details and setup guides, visit the Cloud Provider Observability documentation.

The new Alert rule page features two views, each with their own uses. The Grouped view is arranged by namespaces, so you can drill down to the specific groups you’re looking for from there. The List view is a full list of all the alert rules, designed for quick searches and easy filtering. With the new performance enhancements, you can search and filter instantly.

The new paginated API of the Alert rule page was designed to make alert rule management quick and easy. Whether you want to see a structured hierarchy that illustrates the structure of your alerting setup, or you have thousands of alert rules that you need to search and filter through quickly, the new Alert rule Grouped and List views provide you a simpler, smoother user experience.

On the Alert rule page, click the Try out the new look! button if you haven’t already!

You can now restore dashboards you’ve deleted—no support ticket needed. With the new Recently deleted view, accidentally removing a dashboard no longer means lost work or rebuilding panels from scratch.

When you delete a dashboard, it now moves into a deletion history area instead of being removed permanently. From there, you can browse previously deleted dashboards and restore them to any folder you choose.

Restore dashboards from the new Recently deleted view

Restore dashboards is an admin-only feature. From the Dashboards page, click Recently deleted to see everything that’s been removed in the last 12 months.

Select one or more dashboards to restore. The folder behavior is as follows:

• If a dashboard’s original folder still exists, it’s preselected.
• If multiple dashboards were deleted from the same folder, that folder is preselected.
• If they came from different folders—or the original folder no longer exists—simply choose a new target folder.

Restored dashboards inherit the permissions of the folder you restore them into, making it easy to get them back into the right place.

How deletion history works

Deleted dashboards are stored for up to 12 months, with space for 1,000 dashboards. When that limit is reached, the oldest entries are automatically removed, even if the full retention period hasn’t passed.

Grafana also records who deleted each dashboard and when, giving teams better traceability and confidence.

Availability and limitations

This feature is currently in Cloud only. For more information, refer to the documentation. 

As mentioned above, if you’d like to try it out in our private preview, you can sign up here. And we would love your feedback, you can share it here.

Current limitations:

• Dashboard-level permissions are not restored; restored dashboards inherit folder permissions
• Version history is not preserved (restored dashboards begin at version 1).
• Folders themselves cannot be restored.

The new alert rule page features two views, each with their own uses. The Grouped view is arranged by namespaces, so you can drill down to the specific groups you’re looking for. The List view is a full list of all alert rules, designed for quick searches and easy filtering, which has become much more efficient thanks to recent performance improvements.

The new paginated API of the alert rule page was designed to make alert rule management quick and easy. Whether you want to see a structured hierarchy that illustrates the structure of your alerting setup, or you have thousands of alert rules that you need to search and filter through quickly, the new Grouped and List views provide you with a simpler, smoother user experience.

We have built an entirely new way to visualize and display alerts. This update is designed to significantly simplify the process of triaging firing or pending alerts and enables multiple new ways of exploring your alerts – experience efficient, contextual investigations with this update.

Please note: This new alert visualization is available exclusively for Grafana Managed alert rules.

With the new + Label functionality, you can enrich an alert group with additional context at any point via the UI or API. This makes it easier to:

• Add triage context (for example: triage:needs-follow-up, or status:investigating)
• Support workflows that depend on labels
• Improve discoverability when many teams share the same IRM instance
• Standardize metadata across alerts coming from diverse systems

Highlights:

• User Actions are now GA

  • Additional configuration to manage action state
  • Added an importance attribute (e.g. normal, critical) to categorize user actions
  • Standardized user action event names

• Experimental soft navigation tracking to capture modern Single Page Applications (SPA)

• Support for Web Vitals v5

Faro v2.0 introduces breaking changes from v1. Learn more about the new changes in the Faro changelog and review the migration guide to learn how to upgrade.

Instead of sifting through raw browser events, you can now explicitly label and monitor meaningful actions—like Checkout, Sign Up, or Search. Grafana automatically displays RED metrics for each action and ties them to related traces, HTTP requests, errors, and sessions. This gives teams a complete, intent-driven view of the user experience from the frontend through the backend.

With User Actions, you can:

• Monitor the health of business-critical workflows by seeing which actions aren’t performing as expected.
• Spot slow or error-prone actions before they impact users
• Drill into traces, sessions, and errors for fast root-cause analysis
• Prioritize work by labeling critical actions to focus on actions that has the most impact on your users and teams
• Triage actions instantly in the RCA workbench to quickly identify any underlying service issues impacting your user

User Actions are tightly integrated into the Frontend Observability overview, the action list, action detail pages, and the Grafana Knowledge Graph. They also enable richer insights across your application’s frontend, network, and backend performance.

Start capturing User Actions today using automatic or manual instrumentation via the Faro Web SDK. Learn how to get started in our documentation.

A request to the GetLimits endpoint returns the limits you need to know to keep your fleet running smoothly.

With private connections, you can:

• Reduce your egress costs.
• Improve security by keeping your data within the cloud provider’s network.

Set up your secure connection to Fleet Management today!

The templates provide standardized dashboard layouts designed to help you answer engineering or business questions consistently. For instance, the DORA template allows all teams within an organization to measure delivery performance using a widely adopted industry framework.

All templates include a banner panel that clearly indicates the dashboard is using sample data:

To make the dashboard your own, save it, update each panel’s data source to connect to your own data, and make any other changes to display your data most effectively. In Grafana Cloud, try working with Grafana Assistant to update the dashboard with your data sources and to create queries.

When you’ve finished making your changes, remove the banner panel:

As part of this feature, we’ve also added the Dashboard from template in some “add” menus:

You can access this feature from the New drop-down menu in Dashboards, the quick add button in the top-right corner anywhere in Grafana, as well as the command palette.

Finding the right documentation at the right time can be tricky, especially when you’re in the middle of working in Grafana. The new interactive learning experience, available in public preview in Grafana 12.3, brings help directly into the product, so you don’t have to leave your workflow to find answers.

Accessed through the help icon in the top right corner, interactive learning provides context-aware guidance based on where you are and what you’re doing in Grafana. So whether you’re exploring dashboards, configuring data sources, or building panels, you’ll see relevant documentation, images, and videos right inside Grafana, with interactive tutorials to help you learn by doing.

To try it out, enable the interactiveLearning feature toggle.

The new Kubernetes reader role fixes that. You can grant Kubernetes-specific access to viewers, editors, or anyone without a basic role. Your teams can start using Kubernetes Monitoring while keeping the rest of your environment locked down.

Note: If you previously granted users the None basic role plus plugins.app:access, those users will now lose access to Kubernetes. To restore it, add the new Kubernetes reader role to their custom role.

This new integration automatically forwards incident or alert data to Grafana Assistant for AI-powered analysis and root-cause investigation. Once triggered, the Assistant coordinates multiple specialist agents to identify potential causes, surface insights, and post updates to the related incident timeline and communication channels.

Set up the IRM webhook for Grafana Assistant to start receiving actionable findings and keep your team informed throughout the investigation.

This update is mostly a UX change to make check creation in the UI easier. The new UX also lets us plan for future enhancements to this flow.

The new flow focuses on making it easier to step through the process and addresses some core challenges customers have let us know about.

Specifically:

• Introduction of a right side panel that persists through the entire flow
• Ability to test your check at any step of the flow to validate your check executes as you expect
• Improved the display of logs when testing a check
• Brought Secrets Management into the check creation flow, so you don’t have to navigate to elsewhere to manage secrets

This new design is a foundational step that allows us to continue to iterate and improve the experience in Synthetic Monitoring.

And that’s not all—we’ve also introduced a Source column in the Fleet Management UI, so you can see at a glance where your synced pipelines came from.

Timepoint Explorer visualizes every check execution as an individual event, turning results into a timeline you can explore. Each timepoint represents a probe execution, allowing you to pinpoint when failures happened and easily view the logs tied to those executions.

Explore uptime or reachability at a glance

Toggle between Uptime and Reachability views to investigate results from different perspectives:

Uptime view shows your overall service availability “Can any probes reach my service”

Reachability view breaks down results per probe location, helping you spot region-specific issues faster.

Together, these views make it easier to go from “Alert firing” to “what is failing” and investigating “why is it failing”.

Learn more about Analyzing Results in Synthetic Monitoring

When configuring per-check alerts in Grafana Cloud Synthetic Monitoring, you can now see exactly which notification policies your alerts match,  right from the UI.

The new Alert Routing Summary panel provides a clear, at-a-glance view of:

• The labels automatically included with your alert.
• Which notification policies those labels match.
• The destination contact points where alerts will be delivered.

In the example below, a check labeled label_team=backend matches a backend notification policy, making it clear that alerts will route to the correct destination.

This feature complements per-check alerts and makes it easier to understand, verify, and debug alert delivery.

The PPL query editor now contains a multi-line box which supports syntax highlighting, PPL keywords and commands suggestions, and index fields. To see the new features in action, select PPL in the Query type dropdown of the query editor and start typing. Specify index name(s) in the datasource configuration to narrow down the list of suggested index fields.

We have also added a “Kickstart your query” button to the PPL query editor, which contains useful sample queries. Select a query to copy it to the query editor and edit it further.

For more information on PPL syntax and supported commands, refer to AWS OpenSearch documentation for PPL language.

PPL language support is available from plugin version 2.31.0.

You can now list the log anomalies in your service directly in Grafana. CloudWatch Log Anomaly detection is a new query type for this data source that leverages machine-learning and pattern recognition to establish baselines of typical log content:

You can also use the Logs Insights query editor and the anomaly command to define and display log anomalies ad-hoc. See CloudWatch Logs Insights documentation for more on anomaly detection. 

Log Patterns

Group log entry messages (or other string values) matching specific patterns with the patterns command using any CloudWatch Logs query language. The associated diff command compares the log events found in your requested time period with the log events from a previous time period of equal length, returning results that describe the change.

We have added language support for both commands to CloudWatch Logs Insights language editor.  More info can be found in the AWS Log Patterns documentation and the AWS Logs Insights syntax guide.

During the marketplace review and approval process, we took the opportunity to modernize and streamline our Slack integration. These updates ensure the integration is fully aligned with Slack’s latest platform and security standards, making it faster to install, safer to operate, and easier to trust.

Install the Grafana IRM Slack app from the Marketplace to bring Grafana IRM’s powerful capabilities directly into Slack, where collaboration and communication already happen.

• Explain the data you see in a panel or on a detail page
• Start an investigation for potential issues
• Review the health of Kubernetes objects and resource consumption
• Provide recommendations
• Launch an investigation
• Follow any additional queries or commands you give it

To use the Assistant, click any Assistant icon available on a panel.

Or on any detail page, click the Investigate button.

To get started, add a new variable of type Switch in your dashboard settings. For details, refer to add a switch variable.

This milestone comes after three years of dedicated development, bringing a powerful data type to Prometheus users.

Native histograms improve histogram resolution without requiring users to adjust the buckets. This enables the team to move faster without manual configurations.

The settings available in the drawer are:

• Panel time range - Override the dashboard time range with one specific to the panel.
• Time shift - add a time shift in the panel relative to the dashboard time range or the panel time range, if you’ve set one.
• Time comparison - Compare time series data between two time ranges.
• Hide panel time range - Hide panel time range information in the panel header.

These time settings, when configured, are displayed in the panel header by default:

To access the Panel time settings drawer, click the panel menu and select the Time settings option:

To try out panel time settings, enable the panelTimeSettings and timeComparison feature toggles.

Time comparison

As part of this update, we’ve also added a much-requested feature, the ability to compare data between two time ranges in the same panel:

With this update, you can easily perform time-based (for example, month-over-month) comparative analyses in a single view, on the same panel. This allows you to gain insights essential for trend tracking and performance benchmarking. The addition of this feature has eliminated the need to duplicate panels or dashboards to achieve this type of analysis.

The logs visualization is one of the core visualizations of Grafana, and we’ve rebuilt it from the ground up to not only address performance issues, but also add many new features and display options to innovate in the space and provide the best possible log browsing experience, such as:

Logs highlighting

We added a grammar definition to add colors to different parts of the logs, making them easier to read and facilitating getting important context from their contents.

Font size selection

We added a bigger font size by default, with an option to select a smaller font if you need it.

Client-side search and filtering

We added client-side filter by level and search by string to help you find those logs that you’re looking for.

Timestamp resolution

Logs are displayed with timestamps in milliseconds by default, with an option to use nanosecond precision.

Redesigned Log Details

When you want to know more about a particular log line, accessing its labels, fields, and links, you now have a completely redesigned component with two versions: inline display below the log line, or as a resizable sidebar.

When details are displayed as a sidebar, you can also keep many different details open and quickly toggle between them as tabs.

For both modes, you can now access a sidebar with filters to search for strings withing field names and values, and buttons for the usual actions such as copy log line or link, and explain in Grafana Assistant if it’s available in your Grafana instance.

For Loki, we split fields by type, having dedicated and collapsible sections for indexed labels, strucuted metadata, and parsed fields.

And when a log line contains a trace id, which is configured as a derived field, there is a new panel to render an embedded trace within the details panel.

Redesigned log line menu

You can now find the log line menu as a dropdown menu on the left side of each log line, providing access to Logs Context, toggle Log Details, copy log line, copy link to log line, and Explain in Assistant if available.

Inspired by the analog component in the table visualization, the field selector displays an alphabetically sorted list of fields belonging to all the logs in display, with a percentage value indicating the amount of log lines where a given field is present. From this list, you can select fields to be displayed and change the order of display to the one that better fits your current exploration.

With GCP PSC for Grafana Private DataSource Connect (PDC), you can connect your PDC agents running in your GCP private network to Grafana Cloud while staying on the GCP network.

By connecting your PDC agents via GCP Private Service Connect, you avoid traversing the public internet. This has security benefits and helps lower your network egress costs.

To set up GCP Private Service Connect for PDC Agents, follow the instructions from our public documentation.

Normally when you send telemetry from your GCP network to Grafana Cloud, you incur in network egress fees and your data, though encrypted, traverses the public internet.

With GCP Private Service Connect, you can:

• Reduce your GCP egress costs.
• Improve security by keeping your data within the Google network.

To set up GCP Private Service Connect in Grafana Cloud, follow the instructions from our public documentation.

Use the Export as image option to generate an image:

A preview of the image is displayed. Click Download image:

Using dashboard images, you can quickly send data to your team in chat, added to a presentation, or attached to an email.

The generated image reflects exactly what you’re seeing. If you want to generate the image at a different size or zoomed in or out, simply make these changes in the dashboard or in your browser.

What We’ve Improved

Storage Capacity

We previously had a limitation of about 500 LBAC rules per data source, which was challenging for some customers who needed more rules to meet their requirements. We’ve resolved this by expanding our storage capacity, allowing for significantly more rules to be stored.

Performance Considerations

Query Performance

While the storage concern has been alleviated, query performance can be affected when many LBAC rules are applied. Performance depends on several factors:

• Number of rules applied: Only rules for the teams a user belongs to are applied to queries, not all stored rules
• Rule complexity: Complex rules with many labels can slow down queries
• Datasource type: Some datasources handle LBAC rules more efficiently than others

Technical Limits

There’s a limit on how much data can be sent with each query. Queries exceeding this limit will fail. This becomes relevant with complex rules containing many labels.

How Rules Work

LBAC rules are configured per team. You can store a large number of rules applied to as many teams as you would like. The rules that are actually applied depend on the user running the query and which teams they belong to.

For example, a data source has rules configured for Team A and Team B. A user belonging to Team A but not Team B will only have the rules configured for Team A applied.

This means that the total number of rules saved for a specific datasource doesn’t necessarily affect query performance, but rather the number of rules that apply to the user’s teams.

Best Practices

Performance Guidelines

• Target rule count: In testing, we’ve found that queries generally work well with around 50 LBAC rules applied
• Rule design: Keep rules as simple as possible while meeting your access control requirements
• Team structure: Consider how your team structure affects which rules are applied to each user’s queries

There is no hard limit on the number of LBAC rules that can be sent on a query. This limit will vary depending on the datasource type, the user/team structure, and the rules themselves.

This enhancement allows users to easily query and visualize datasets registered in Unity Catalog while maintaining the same fine-grained permissions and lineage tracking defined in Databricks. With this integration, teams can build dashboards confidently knowing their data follows centralized governance, access control, and compliance standards across all workspaces.

To enable support for Databricks Unity Catalog, you must check the corresponding checkbox in the configuration page of your Grafana Databricks plugin:

If the checkbox is left unchecked, the old layout with dataset and table dropdown menus is displayed.

This feature is in public preview and is only available on Grafana Enterprise and Cloud accounts.

SolarWinds is a network and infrastructure monitoring platform that provides tools to monitor, manage network performance, server health, and application availability. 

The SolarWinds data source plugin enables you to query and visualize key infrastructure metrics such as CPU Load, Memory usage and Active Alerts directly within Grafana dashboards. This integration allows teams to correlate infrastructure health with application performance, enabling proactive monitoring and faster incident response across hybrid and multi-cloud environments.

Available in public preview in Grafana Cloud and all supported versions of self-managed Grafana Enterprise.

The Batch Visibility Gap

Managing critical Jobs and CronJobs across multiple Kubernetes clusters has historically been a source of operational friction, leading to:

• Undetected Failures: Missed execution windows and silent job failures that are only discovered hours later.
• Prolonged MTTR: Debugging requires manual, multi-command context-switching across different logs, metrics, and event streams.
• Inefficient Resource Use: Lack of historical resource data makes it difficult to right-size requests, leading to either wasted cloud spend or resource contention.

Key outcomes of the new Job Monitoring feature

This new feature delivers essential insights and consolidated diagnostic tools, ensuring the reliability and cost-efficiency of your scheduled tasks.

1. Guaranteed Execution Integrity

• Centralized Health View: The All jobs list provides a single source of truth across all Clusters and Namespaces, with color-coded status indicators for immediate failure detection.
• Punctuality Confirmed: For CronJobs, the feature explicitly compares Last scheduled against Last succeeded, instantly flagging any skipped runs or execution gaps that could impact data pipelines or maintenance routines.

2. Accelerated Diagnostics and Root Cause Analysis

• Consolidated Detail Page: Clicking a job name opens an integrated detail view, consolidating the Pod status, logs, and events—eliminating the need for manual kubectl commands.
• Pattern Identification: The historical Runs table tracks duration and success rates over time, enabling teams to spot flaky jobs, analyze trend data, and drastically reduce Mean Time to Resolution (MTTR).

3. Optimized Performance and Cost

• Resource Right-Sizing: Dedicated views for CPU and memory usage (including trends) facilitate precise resource provisioning. This data empowers you to identify wasteful over-provisioning and detect performance-impacting issues like gradual memory leaks.

Action: Access the new capabilities by navigating to the main menu and selecting All jobs within the Grafana Cloud Kubernetes Monitoring App.

Click to mark your most-used data sources as favorites, and they’ll appear at the top of the list every time – no more scrolling or hunting through endless options.

Log context is a feature that displays additional lines of logs surrounding a log entry that matches a specific search query. This helps in understanding the context of the log entry, analog to the -C parameter in the grep command.

The new logs context component takes advantage of all the new options and capabilities of the new visualization, and introduces a new option to select the specific amount of time before and after the referenced log line, which ranges from a hundred milliseconds up to two hours.

Note that this feature is experimental. To try it out, enable the newLogContext feature toggle.

Seamlessly synchronize configuration pipelines from external sources like Git and Terraform, ensuring your remote configurations stay updated across environments. Automate pipeline creation, updates, and deletions to minimize manual effort. Whether you’re managing a handful or a hundred pipelines, Fleet Management handles batch updates efficiently while maintaining data consistency. Learn more in our set up documentation.

Bonus: pipeline syncing is just the beginning! We’re working on a GitHub Action to take your GitOps workflow to the next level. Stay tuned!

With Grafana Assistant you no longer need to juggle tabs or write complex queries when incidents start. Open Grafana, click the sparkle icon, describe the goal, and let the Assistant guide you. It runs inside your Grafana session, inherits your permissions, and grounds every answer in your own telemetry so you stay in control from first question to resolution.

GA means the Assistant is ready for everyday use across your teams. It helps you write and validate queries, build and refine dashboards, navigate to the right places in Grafana, and explain patterns in plain language.

What this means for you

Faster triage and root cause. Ask for a goal. Get queries, context, and a clear next step. Use investigations for parallel analysis when the issue spans services.

Less query toil. Describe intent. The Assistant drafts, validates, and optimizes queries before you commit them to dashboards or alerts.

Cleaner handoffs. Copy summaries to tickets or chat. Share links to dashboards and investigations. Use feedback controls to improve results over time.

Safe, secure, and trustworthy. RBAC controls access. Requests route through Grafana Cloud with encryption, auditing, and least-privilege design. No business data training of external models. Providers are configured for zero retention.

Key highlights you will use on day one

Grafana Assistant workflows support multiple audiences: site reliability engineers and operators who triage incidents, dashboard authors and analysts who publish insights, and platform administrators who coordinate access and automation.

Start with your own voice and let the Assistant handle the mechanics. When you describe a goal, it drafts PromQL, LogQL, SQL, and more, validates the syntax, and suggests safer, faster alternatives before you ship changes to dashboards or alerts.

Full signal coverage. Work across metrics, logs, traces, and profiles in one flow. Use Prometheus, Loki, Tempo, and Pyroscope without switching tools. Pivot between signals during incidents.

SQL data sources. Blend business and ops data. Query PostgreSQL, MySQL, ClickHouse, or BigQuery and correlate KPIs with system behavior.

Knowledge Graph. If your stack uses the Knowledge Graph, the Assistant surfaces related context.

Infrastructure memory. Give the Assistant live service context. Run discovery to capture services, dependencies, labels, and triage checklists. Answers and investigations stay aligned to your topology. Refresh after changes.

Rules. Ship a consistent, governed Assistant across your organization. Set tone, encode best practices, add infra hints, and nudge workflow escalations. Apply per user or tenant. Update any time.

MCP servers. Bring your tools and knowledge into the Grafana Assistant. Connect MCP servers with HTTPS and OAuth 2.0 with PKCE. Control scope with RBAC. Remove access when no longer needed.

Get started right away

Administrators can enable Grafana Assistant for your Grafana Cloud account by accepting its terms and conditions in the plugin configuration. Once enabled, open the sidebar and ask a focused question, adding an @ mention for a data source or dashboard to ground the answer. The built-in tutorial and prompting guide help new users ramp up quickly.

Grafana Assistant is free to use until January 1, 2026, with fair-usage allowances and built-in usage indicators to help you track adoption.

Read our blog or watch a short video on how you can get started with Grafana Assistant.

Assistant Investigations gives you a faster path from “what is happening” to “what to do next.” When an incident spans services or signals, you switch the Assistant to Investigation mode, describe the problem, and let specialist agents explore metrics, logs, traces, profiles, and even SQL in parallel. The workflow runs inside Grafana Cloud, respects your permissions, and keeps findings grounded in your telemetry.

What are investigations?

Investigations coordinate multiple agents at once. A lead investigator plans the work while Prometheus, Loki, Tempo, and Pyroscope specialists gather evidence. A reporter compiles results into a structured report you can trust. You get a live activity feed during the run, then a finished artifact with a Summary, a detailed Report, a Timeline, and an Activity log.

The experience is conversational. You steer with plain language, add @ mentions for services or data sources, and correct scope as you learn. Use thumbs up or thumbs down to nudge agents toward the right paths. The more context you provide, the better the result.

Context carries through every step. Discovery-powered infrastructure memories keep service names, dependencies, and label hints current so agents start with the right map. Mentions pull that context into the investigation automatically, which reduces guesswork and keeps the findings aligned with your environment.

You choose when to escalate. Stay in chat for quick checks. Launch an investigation when you need breadth or a durable write-up. Agents fan out across Prometheus, Loki, Tempo, Pyroscope, and SQL data sources, then return a focused narrative you can paste into incident channels or tickets.

Get started with Assistant Investigations

Enable Investigations for your organization, grant the Assistant Investigation User role where needed, then open the Assistant and select Deep Investigation. Describe the symptom, timeframe, and scope. Add @ mentions to anchor data sources or dashboards. You can monitor progress live and request adjustments at any time.

What to try first: pick a recent alert, launch a Deep Investigation, and ask, “Why did checkout latency spike after 17:30?” Watch agents collect evidence, read the Summary, then use the Timeline to brief your team. If the scope changes, add a new hint or mention the service you care about and keep going.

Now in public preview

Security and privacy follow Grafana Cloud standards. The Assistant operates within your existing RBAC, routes calls through Grafana Cloud, and encrypts data in transit and at rest. Your organization’s data is not used to train external models. Sub-processors are configured for zero retention, and usage is logged for auditing and quality.

Assistant Investigations is a public preview. Support is best effort and changes may occur before GA.

The feature is free to use until January 1, 2026, with fair-usage allowances that include a soft limit of 50 investigations per user per month.

Once enabled for your organization, you can chat with the Assistant from Slack to draft PromQL, LogQL, or TraceQL, navigate to the right place in Grafana, or kick off deeper analysis that you review inside Grafana.

Sign up for access

Once your access is approved, you will receive setup instructions. After enablement, you can start a conversation with the Assistant in Slack, get answers in plain language, and jump into Grafana for richer views or to launch deeper analysis when needed.

This is a private preview and features may change before GA.

That’s why we’ve added a “Notification History” feature to the IRM Mobile App, so you can track down the notifications you’ve received and review important events in chronological order.

On the home screen, you’ll see a new icon in the top-right corner. Tap it to view your recent notifications from the last 30 days, including alert groups, shift swap requests, and upcoming on-call shifts.

We hope this feature improves on-call engineers’ quality of life!

We’re excited to announce the general availability of the Cost Management and Billing app for Grafana Cloud! This redesigned experience provides a central place to manage all your Grafana Cloud usage and spend.

Key features include:

• Cost attribution: Break down costs by label (team, project, etc.) for better accountability
• Usage-based alerts: Get notified when usage exceeds thresholds with easy to set up alerts
• Invoice management: View and download monthly invoices directly in the app
• Spend insights: Compare month-over-month spending trends to identify changes and optimize your spend

To learn more about the full functionality, view our documentation.

Note: With this release, the Cost Management Hub has been deprecated. All functionality previously available there is now accessible through the navigation bar, within the Cost Management and Billing app, or in Adaptive Telemetry.

• apikeys:read
• apikeys:create
• apikeys:delete

The following roles were also removed:

• fixed:apikeys:writer
• fixed:apikeys:reader

These actions and roles no longer had any effect in Grafana, so this change does not have any user impact.

Operators deploying custom access control roles or customized basic role via Terraform will need to update their role definitions to remove the affected actions.

What does this mean for you?

• Unrestricted Querying: Gain full access to all Honeycomb API operations, including complex filters, aggregations, transformations, breakdowns, and more.
• Deeper Insights: Craft highly specific and nuanced queries to extract the exact data you need, leading to more profound insights into your application performance and system behavior.

We’ve also introduced a new Variable Query Type. Traditionally, queries relied on the dataset slug, which is a transformed, lowercase version of the dataset name. However, since the slug and display name don’t always match, the dataset name can be just as valuable when building queries. To address this, we’ve added support for selecting datasets by either slug or name, giving users more flexibility and clarity when working with their data.

How to get started?

Simply select the “Raw Query” option from the query type dropdown in the Honeycomb query builder. A JSON text area will appear, allowing you to input your queries in the same format accepted by the Honeycomb API. For guidance on supported operations and syntax, refer to the official Honeycomb API documentation’s “create query” page.

To learn more, refer to the following video demo:

Raw Query requires you to enable the honeycombEnableRawQuery feature flag. If you are a Grafana Cloud customer, contact support to request access and let us know what you think!

Internationalization of plugins is here! You can now make your plugin more user-friendly by translating it into multiple languages. You can find the list of available languages in GitHub, and  how to translate plugins in our docs page. For Microsoft users, we’ve added translations to MS SQL and Azure Monitor data sources.

The k6 Operator offers two Custom Resource Definitions (CRDs): TestRun, to execute a distributed k6 test on a Kubernetes cluster, and PrivateLoadZone, to create a load zone on a Kubernetes cluster for Grafana Cloud k6 to run k6 tests in.

With the v1.0 release, we are committing to a more predictable release schedule, maintenance approach, and a stable upgrade path for the k6 Operator. This includes:

• Semantic Versioning: The k6 Operator will now follow Semantic Versioning 2.0. In general, breaking changes will only be introduced in major releases, with prior deprecation warnings. Refer to stability guarantees for details.
• Predictable Release Schedule: We are committing to a new minor version release every 8 weeks, with more frequent patch releases for bug fixes.
• New Upgrade Guide: To assist our users, we have added a guide on how to upgrade the k6 Operator.

For more details, refer to the release blog post.

While you can turn on this option under Cell options and have it applied to all cells in the table, it’s typically used as an override on a sub-set of cells instead.

Benefits you get right away:

• Stronger security: No more “forever” tokens left lingering in your environment.
• Policy alignment: Every token follows the lifetime you define, supporting your compliance requirements.
• Easier audits: Shorter-lived tokens reduce the risk of stale credentials being overlooked during reviews.
• Smaller attack surface: Limiting token lifetimes lowers the chance of long-term misuse if a token is ever exposed.

Head over to your Cloud Portal settings to set your organization’s maximum token expiration and keep your environments safer by default.

We recently announced SCIM for user and group provisioning in Grafana, and now we’re excited to announce that SCIM configuration is easy with the SCIM Configuration UI. Here you can easily toggle SCIM configuration options without restarting your instance, find information like your Tenant URL, Domain, and Stack ID for configuring SCIM in your IdP, and learn how to create a service account and token with the correct permissions.

The reject_non_provisioned_users setting in Grafana defaults to false. When enabled, non-SCIM provisioned users are prevented from signing in. Cloud Portal users can always sign in regardless of this setting.

From Dashboards and Explore, save custom queries:

Populate the existing query editor with a saved query:

Or add a subsequent query using a previously saved query:

In the Saved queries drawer, you can edit, search, and filter to find saved queries. Click the star next to a query to easily access it in the Favorites tab. You can also set queries to be shared with others in your organization:

Lastly, you can add queries you’ve run in the past two weeks to your saved queries from the Query history tab in Explore.

This feature is in public preview and is only available on Grafana Enterprise and Cloud accounts. This feature will gradually roll out to all Grafana Cloud users over the next few days with no action required. To try out this feature on Grafana Enterprise, enable the queryLibrary feature toggle.

Source maps are files that connect your minified, bundled production code back to its original, unminified source files. Without a source map, an error trace is less readable and unable to point to a source. You can either host Source Maps publicly or use our upload source map functionality privately in Grafana to ensure your traces are readable.

What’s new:

• Search for source maps by name.
• Distinguish source map bundles with new metadata showing Created and Last Updated timestamps
• Delete individual or many source maps at once. This is useful for clearing out old or unused files or when you need to re-upload a similar bundle.

Source Maps are public preview and be accessed by going to your Frontend Observability App -> Settings -> Source Maps.

• Cell types, Pill and Markdown + HTML
• Wrap header text
• Frozen columns - Freeze as many columns as you want, starting from the left side of the table.
• Max row height - This is helpful when you apply text wrapping to one or more columns, which can significantly increase the height of a cell.
• Tooltip from field - Render the content from one cell as the tooltip for another cell.

We’ve also added new functionality to the table footer, significantly improved its display, and streamlined the configuration process.

Lastly, we’ve moved Wrap text from the cell option level to the table option level.

New cell types

We’ve added two exciting new cell types: Pill and Markdown + HTML

Pill

The Pill cell type displays each item in a comma-separated string in a colored block:

The colors applied to each piece of text are maintained throughout the table. For example, if the word “test” is first displayed in a red pill, it will always be displayed in a red pill.

The pill cell type supports the following data formats:

• Comma-separated values (cows,chickens,goats)
• JSON arrays of uniform ((["cows","chickens","goats"])) or mixed ([1,2,3,"foo",42,"bar"]) types

Markdown + HTML

The Markdown + HTML cell type displays rich Markdown or HTML content, rendered using the GitHub-Flavored Markdown spec. This is useful if you need to display customized, pre-formatted information alongside tabular data, such as formatted strings, lists of links, or other dynamic cases.

Tooltip from field

With the tooltip from field feature, you can use any of the fields in the table as the source of tooltip content.

For more information, refer to Generate tooltips from table fields.

Updated footer

The table footer displays the results of calculations (and reducer functions) on fields. We’ve updated the footer so you can apply multiple calculations or reducers simultaneously.

For more information, refer to Improved footer for table.

We’ve improved the general experience by providing a more intuitive editor that supports better formatting, quick copying, easy expansion, and a shortcut to instantly run your queries.

We’ve also implemented support for autocomplete. We’ve additionally added column/field autocomplete, available after enabling the sqlExpressionsColumnAutoComplete feature toggle, which is provided on an experimental basis.

Lastly, we know the real challenge isn’t just writing SQL — it’s asking the right questions and understanding complex queries. That’s why we’ve integrated Grafana’s LLM directly into the editor. You can now generate SQL queries from natural language and get instant explanations of existing queries, helping you move from idea to insight faster than ever.

Give it a try and let us know how you like it!

This is enabled by two key enhancements:

• Dashboard data source support - You can now define ad hoc filters in dashboards that use the -- Dashboard -- data source.
• Set filters from the bar chart tooltip - Previously, setting ad hoc filter values was only possible on table visualizations. Now, you can do this from the bar chart as well.

Together, these updates unlock a flexible workflow: a single source panel defines the dataset, and ad hoc filters make it easy to explore and refine.

Watch the video to see these enhancements in action:

The key changes are support for filtering and the ability to browse recent resources.

You can now filter the resources displayed in the picker based on their subscription, type, and location, allowing you to easily browse your resources and swiftly find those that you’re interested in.

Additionally, once you’ve found those resources, the recent resources tab will support you in discovering them again if needed, rather than having to conduct your search again.

To enable the enhanced resource picker enable the azureResourcePickerUpdates feature toggle today.

In the panel editor, you can see that the Calculation option now allows you to select multiple calculations at once:

We’ve also removed the Show footer switch in the panel editor. Now, the footer is only displayed if you select an option in the Calculation drop-down list. With this change, you can never have an empty footer by mistake.

The Fields option has also been removed, and by default, footer calculations now apply to all fields in the table. If you want the calculation to only apply to a subset of fields, you can control that using field overrides. This aligns with other table options and provides a more consistent user experience.

With this update, we’ve also improved non-math reducer functions (like First and Last) so that they’re applied to non-numeric fields, like strings. Previously, this functionality wasn’t supported for non-numeric fields.

Note that calculations applied to cell types like Markdown + HTML might have unexpected results.

Previously, actions could only make direct HTTP requests from the browser, which are subject to CORS restrictions, significantly limiting their usefulness.

With this update, you can choose between:

• Direct from browser - Unauthenticated requests
• Infinity data source - Routing requests through a preconfigured Infinity data source

To try out this feature, enable the vizActionsAuth feature toggle. Note that users with the Viewer role can’t see or execute actions.

Normally when you send telemetry from your GCP network to Grafana Cloud, you incur in network egress fees and your data, though encrypted, traverses the public internet.

With GCP Private Service Connect, you can:

• Reduce your GCP egress costs.
• Improve security by keeping your data within the Google network.

To set up GCP Private Service Connect in Grafana Cloud, follow the instructions from our public documentation.

The Logs Drilldown JSON viewer provides you a log line viewer specifically created to make it easier to read and understand your JSON log data.

The Logs Drilldown JSON viewer displays JSON log data in a structured and collapsible view that lets you focus on specific log lines or the fields and values in your JSON logs. You can quickly filter for specific information with a single click, and make related visualizations more relevant and focused by including or excluding specific fields and values.

We hope these new capabilities will improve your ability to debug and analyze your log data, helping you find the “signal in the noise” and gain deeper insights.

We started the rollout in May. Since then, we’ve gathered feedback, measured impact, and improved testing. We’re now confident the APIs are stable and have made them the default.

The new APIs make Git Sync possible, and are the key change that powers our new Terraform provider and the creation of the GrafanaCTL CLI tool. Currently, only dashboards, folders, and a few other elements, such as announcement banners and playlists, use this new model; however, we plan to expand it to all Grafana resources.

For more information, refer to the documentation.

You can now create alerts for each check in your Synthetic Monitoring application. For example, you can create an alert based on the number of check failures in a specific time window, with different settings for each one of your checks.

Additionally, there are specific alerts available for certain check types:

• For HTTP and TCP checks, you can create an alert if a TLS certificate is set to expire within a certain number of days.

• And for HTTP, DNS, and Ping checks, you can create alerts based on latency.

The legacy alert rules based on sensitivity thresholds can still be used in combination with per-check alert rules.

Per-check alert rules give users more granular control over their alerting needs, making it easier for users to configure the right alert threshold depending on the services they’re monitoring and the check settings. These alerts are powered by Grafana Managed Alerts.

To learn more, refer to the per-check alerts docs.

The feature integrates seamlessly into the incident detail view, displaying the latest key status update prominently while providing easy access to the complete history. When creating new updates, you’ll see selectors for incident status and severity changes, ensuring that critical state changes are communicated clearly to all stakeholders.

This enhancement helps you keep your incident response team aligned and informed, whether you’re responding from a mobile device or browser.

CloudWatch metric streams with Amazon Data Firehose offer you a less expensive, highly accurate way to import metrics from AWS into Grafana Cloud. As with CloudWatch metrics scrape, CloudWatch metric streams do not require you to deploy or maintain Alloy agents or servers and can be configured as code using Terraform or using a CloudFormation template.

CloudWatch metric streams capture metrics in near real-time, without sampling and little delay, offering you a more complete and accurate dataset compared to metrics scrape, which can miss spikes or drop metrics during polling intervals. Streams push data with less lag, which improves responsiveness for alerting, dashboards, and downstream processing. Furthermore, streams can be up to 5x cheaper than scrapes, since you are not paying for compute resources to pull and process metrics. You pay only for what you stream, reducing overhead.

To setup CloudWatch metric streams:

1. Go to your Grafana Cloud portal.
2. Expand Observability and Cloud provider, then select AWS in the main menu.
3. Select the Configuration tab.
4. Select the CloudWatch metric streams card.

For more information about configuring CloudWatch metric streams as code, refer to Configure metric streams with Terraform in the documentation.

This is especially useful for visualizing real-time data flows, system states, or transitions, where directionality changes based on metrics.

Watch the video to see these enhancements in action:

When you toggle the switch on, you can select from a drop-down list any of the fields in the table to be used as the source of the tooltip content. All table fields are available for selection, whether visible or hidden.

While you can turn on this option for all the cells in a table under Cell options, it’s most helpful when used as an override on a subset of cells instead.

For example, the following table has five visible fields (columns) as well as a hidden field called “Info”. The values from the “Info” field are used as tooltip text for the “Short text” cells using the Cell options > Tooltip from field override property:

The content of the tooltip is determined by the values of the source field, and you can use field overrides on the source field to manipulate the display of that value.

For more detailed information, refer to the documentation.

You can now seamlessly use extensions from both Grafana Cloud and the k6 CLI:

• Local execution: k6 run test.js
• Cloud execution: k6 cloud run test.js
• Local execution, streaming test results: k6 cloud run test.js --local-execution

You can try this test.js example script that uses the k6/x/faker extension:

js Copy

import faker from 'k6/x/faker';

export const options = {
  cloud: {
    name: 'Extensions in Grafana Cloud k6',
  },
};

export default function () {
  console.log(faker.person.firstName());
}

On top of it, you can also pin to specific versions of k6 and extensions in your script. For cloud-executed tests, this is only available for contracted customers. The previous test.js example can be updated to:

js Copy

'use k6 = 1.0.0';
'use k6 with k6/x/faker <= 0.4.3';

import faker from 'k6/x/faker';

export const options = {
  cloud: {
    name: 'Extensions in Grafana Cloud k6 (Version overrides)',
  },
};

export default function () {
  console.log(faker.person.firstName());
}

Taking these examples to action:

Grafana Assistant is an AI tool that can…

• … create dashboards.
• … write queries.
• … help you get onboarded.
• … support you in learning Grafana.
• … support in investigations and incidents.

Read our blog or watch a short video on how you can get started with Grafana Assistant.

This enterprise plugin includes two built-in dashboards to help you quickly get started visualizing Jenkins data:

• Jenkins Overview dashboard provides an overview of the Jenkins instance, including all projects, nodes, executor status and build queue.
• Jenkins DORA Metrics dashboard provides information on four key metrics used to assess software development team performance: deployment frequency, lead time for changes, change failure rate and time to restore service. These metrics focus on both the speed and stability of software delivery.

Available in public preview in Grafana Cloud and all supported versions of self-managed Grafana Enterprise.

Secrets management gives you a centralized place to securely store sensitive data like API keys, passwords, and tokens.

You can reference secrets by name in your scripted and browser Synthetics checks so that sensitive values are never visible or stored as cleartext. Secret values are also automatically masked in all Synthetic Monitoring output to prevent accidental exposure.

Access to create, update, and delete secrets is currently restricted to admin users. However, anyone with permissions to edit checks can use secrets provided they know the secret name.

Currently secrets can only be accessed in scripted and browser checks, but support for more check types will be available in the future.

To learn more, refer to the secrets management documentation.

This integration opens up new possibilities for interacting with tracing data. You can now connect LLM-powered tools like Claude Code or Cursor to Grafana Cloud Traces, enabling you to:

• Explore services and understand interactions: LLMs can be used to teach new developers about service interactions within an application by analyzing tracing data. For instance, a new developer could ask the AI to explain how their services interact. The AI would use live tracing data from Cloud Traces to answer these questions.
• Diagnose and investigate errors: You can leverage LLMs to identify and diagnose errors in your systems. The AI can answer questions like “Are there errors in my services?”, “What endpoints are being impacted”, etc?
• Optimize performance and reduce latency: LLMs can assist in identifying the causes of latency and guiding optimization efforts. By analyzing trace data, an LLM can summarize operations in a request path, pinpoint bottlenecks, and even suggest code changes to improve performance, such as parallelizing operations.

Setting up an LLM agent with Grafana Cloud Traces requires a Grafana Cloud API token and configuration. This capability is under active development and we welcome early adopters to try it out and share feedback.

With this update, you can:

• Customize webhook requests with support for any HTTP method
• Use templated URLs, headers, and request bodies
• Dynamically reference incident data and prior webhook responses
• View webhook execution details directly from the incident or alert group timeline
• Manage incident-related webhook configurations using Terraform

The legacy Incident Outgoing Webhooks app and existing configurations will continue to work. However, we recommend migrating to the shared IRM webhook experience to take advantage of the latest features and ensure long-term support.

The Challenge:

Creating effective alert rules often requires knowing exactly what query to write upfront. But sometimes you need to explore your data first — understanding how metrics behave across different label combinations, spotting patterns, and building confidence in your monitoring approach before committing to an alert rule.

The Solution:

Metrics Drilldown integration brings visual data exploration directly into alert creation. Launch from your alert rule editor into Grafana’s queryless Metrics Drilldown experience, where you can “quickly find related metrics with just a few simple clicks, without needing to write a PromQL query.” Explore your Prometheus data through intuitive point-and-click interactions, visualize how metrics break down across labels, and leverage “smart segmentation” with “optimal chart types selected automatically.” When you discover the perfect visualization, click through to Explore to grab the underlying query and use it in your alert rule.

What Makes This Special:

Exploration-first workflow - See your data before writing queries

Queryless integration from alert editor → Metrics Drilldown → Explore → back to alerts

Visual confidence building-understand exactly how your metrics behave

Try It Now:

Access Metrics Drilldown from any Prometheus alert rule query editor in Grafana 12.1.1 and later.

The Result:

Transform from “I think I need to monitor this” to “I know exactly what to alert on and why” - all through visual exploration that makes alert creation both intuitive and confident.

Tailscale allows you to create a secure network (called a tailnet) by directly connecting users, devices, and resources. This new integration adds an ephemeral machine to your tailnet on your behalf. You can add tags to these machines, which allow you to configure Tailscale ACLs and Grants, giving you full control of what your Grafana Cloud stack can access.

Check out the video below to learn more about how it works.

Key benefits of the integration

• Secure and private data access: With Tailscale, querying private data sources can remain private — no endpoints need to be opened to the public internet.
• Simplified setup and reduced operational overhead: Private Data Source Connect with Tailscale is a simpler alternative to running PDC agents. To connect to a data source on a tailnet, you simply need to enter your data source’s tailnet address or machine name using MagicDNS and provide a Tailscale auth key.
• Extended observability with Grafana Cloud: Grafana Cloud offers advanced observability solutions and greater convenience than a self-hosted stack. You can query data sources in private networks from anywhere, making it ideal for distributed teams and remote work.

Sign up today

This integration is in private preview, so for now, we recommend using it only for non-production stacks. You can sign up for the private preview, and we will contact eligible users for onboarding.

The canvas visualization editor now offers a completely re-engineered pan and zoom experience.

You can now place elements anywhere—even beyond panel edges—without disrupting connections or layouts. Background images stay consistent, connection anchors rotate with elements, and an optional Zoom to content toggle automatically fits your canvas content to any view. Constraints remain intact thanks to a transparent root container, ensuring layout behavior stays reliable across pan and zoom operations.

Constraint system support

The improved pan and zoom works seamlessly with canvas’ constraint system. An invisible root container ensures all elements maintain their relative positions and constraints, even during heavy navigation or scaling.

Stable connections

Connection anchors stay exactly where you placed them, even during rotation or resizing. This prevents misalignment and ensures your diagrams remain structurally accurate.

Zoom-to-content toggle

With one click, you can fit your entire dashboard into view. This is especially helpful when switching between large screens and smaller devices, ensuring all elements are visible without manual panning or resizing.

To try out this feature, enable the canvasPanelPanZoom feature toggle.

Watch the video to see these enhancements in action:

• Users looking to export an existing SLO into HCL format should locate their SLO on the Manage SLO screen, click the More drop-down, and select Export
• Users looking to create a new SLO to be exported should navigate to the Manage SLO screen, click the more drop-down on the top right of the page, and select New SLO for export

Additionally, you can now use the Disable for one-click elements option to selectively hide tooltips on elements that have one-click functionality enabled. This enhancement prevents tooltips from interfering with one-click interactions while still allowing tooltips on other elements.

Note that tooltips are only displayed when inline editing is disabled for a canvas.

Using our simple workflows you can enable and configure alerts based on your web app’s errors and web vital metrics. Find and troubleshoot issues sooner now that alert configuration and rules are handled automatically. Additionally, these alerts can serve as templates for you to expand the alerting coverage of your frontend apps. 

Check out the docs to learn more.

Status updates are structured messages that communicate key information throughout the incident lifecycle. Whether you’re confirming impact, escalating to another team, or resolving the issue, use status updates to help track and share progress in a consistent, high-signal way.

You can add status updates directly in the Grafana UI or by using the following Slack commands in the associated incident channel:

• Use /grafana incident update add <content> to post a new status update from Slack.
• Use /grafana incident update to view the latest status update.

To get started, open any incident in Grafana IRM and locate the Status updates section in the activity timeline. Learn more in the docs.

It performs a series of periodic checks against your Grafana instance to highlight issues requiring the server administrator’s attention.

In this early release, it includes checks for plugins, data source connections, and SSO settings.

Learn more and see the advisor in action in the video below:

We welcome any feedback and suggestions, which can be provided through this form.

The advisor is currently in public preview in Grafana Cloud, and private preview in Grafana OSS and Enterprise as of v12.1.

See examples of this transformation on Grafana Play.

• Which API endpoints or assets are causing slowdowns or errors?
• How widespread is a specific performance issue?
• What are the biggest opportunities for optimization based on how issues occur over time.

To get started, log in to your Grafana Cloud Frontend Observability application and navigate to the new HTTP tab. Your real user HTTP data is automatically available for you. No extra setup required. For more information visit our documentation.

View GPU utilization panels on the GPU tabs of Cluster and Node detail pages to discover whether Nvidia GPUs inside the Cluster are appropriately utilized and whether workloads are getting and using the GPU resources that have been made available.

The new sound collection features aircraft alarms, emergency warnings, fire alarms, loud buzzers, and severe warning alarms - each available in both constant and fade-in variations. These sounds were carefully selected based on customer feedback requesting more effective notification options that ensure alerts are never missed.

You can find these new alarm sounds in the mobile app’s sound settings, alongside the existing natural sounds, short tones, and musical options. This gives on-call teams complete flexibility to customize their notification experience based on personal preferences and operational requirements.

The enhanced sound library ensures that critical alerts reach on-call engineers reliably, helping reduce response times and improve incident resolution - because the most sophisticated alerting system is only as effective as its ability to wake up the person who needs to respond.

Imagine you have a spreadsheet with sales data for different regions. With template variables, you can add a drop-down menu to your dashboard that lets users select a specific region. The dashboard automatically updates to show only the data for that region.

What’s new?

1. Variable interpolation: You can now use variables within other variables. This means you can create more complex and flexible variable relationships. For example, you could have a variable for “Country” and another for “City,” where the available “City” options change based on the selected “Country.”
2. Filtering by value: You can create new variables based on the values of existing ones. This allows you to refine your data selections. For instance, you could have a variable showing “Product Categories,” and then create a new variable that only displays “Products” within the selected category.

These new features make it much easier to build flexible, reusable dashboards using data from Google Sheets, giving your users more control and customization.

For more info, check out the Google Sheets variable documentation.

For more details on configuring Entra ID (formerly known as Azure AD) authentication, please refer to the official documentation.

This feature is especially useful when actions require user-supplied input upon being triggered. For example, it enables scenarios like setting a thermostat to a specific temperature or entering a problem description when creating a support ticket.

We continue to celebrate and ensure our opinionated experiences can benefit from our best-of-breed signal exploration workflows by making them embeddable in other contexts.

What’s new?

Instead of being greeted with a basic logs panel and a daunting query interface, you get the full Grafana Logs Drilldown experience right inside the KPI drawer and can more comfortably work with your log lines.

Bonus: Linkable attributes

To create a fully connected experience, we’ve added hooks to log labels and trace attributes so you can jump directly to richer Grafana Cloud solutions in the context of the signals you were exploring. These prefabricated shortcuts can take you straight into Application Observability – our first integration – from logs to a specific pod in Kubernetes monitoring, or from a trace to a specific query in Database Observability.

Enable auto-apply using the Adaptive Metrics GUI, API, or Terraform provider. Auto-apply also optionally works with segmentation; sub-divide your recommendations to manage on a per-team basis, and choose to auto-apply at a segment level.

The feature is currently in public preview. We recommend that you use the Exemptions feature before turning on auto-apply, to ensure any critical metrics are not aggregated inadvertently.

Thanks to Drilldown, you can now spot which dimensions are misbehaving – easily – no more wading through endless trace lists or wrestling with queries. That means dramatically faster MTTR when you need traces to pinpoint root causes!

What’s new?

Now, instead of being greeted with a bare list of traces and a daunting query interface, you get the full Traces Drilldown experience right inside Asserts without jumping between tools or writing a query. Plus, Traces Drilldown automatically inherits the added context that Asserts provides, including pre-selected filters and tabs, so you dive straight into the data that matters.

You can quickly choose your metric, compare trends between attributes, and pick the traces you want to analyze directly within the Asserts application. When you drill into an individual trace, it now expands to the full pane, making the most of your available screen real estate.

One more thing: linkable attributes

To create a fully connected experience, we’ve added hooks to log labels and trace attributes so you can jump directly to richer Grafana Cloud solutions in the context of the signals you were exploring. These prefabricated shortcuts can take you straight into Application Observability – our first integration – from logs to a specific pod in Kubernetes monitoring, or from a trace to a specific query in Database Observability.

With per stream export, you can configure Grafana Cloud to export select log streams to your object storage bucket in AWS, Google Cloud, or Azure. By only exporting a subset of your logs, you can significantly reduce your object storage costs and minimize logs needed to sort through after being stored.

The per stream export experience is generally available now in the Cloud Logs Export plugin.

With AWS PrivateLink, you can:

• Reduce your AWS egress costs.
• Improve security by keeping your data within the Amazon network.

Learn how to set up your secure connection today!

By default, inactive collectors are hidden from view in the Fleet Management application. You can switch the filter to Inactive to see them.

Key updates include:

• Creating reports from dashboards while maintaining dashboard context
• One-page report creation
• Easy access to organization report settings

When you click Share > Schedule report from a dashboard, a drawer opens where you can complete the entire report form. With this change, you’re no longer taken out of the dashboard, but instead maintain your original context. If other reports include this dashboard, they are also listed in the drawer, helping to reduce duplication of efforts.

With this update, we’ve also removed the report wizard and replaced it with a one-page form. The page has sections that can be expanded and collapsed, allowing you to easily access all report options on a single page.

In addition to these changes to report creation, if you’re in Reporting, you now access report settings in a drawer that allows you to maintain your context. (If you’re in Dashboards, clicking Report settings takes you to the Reporting feature in another browser tab, where the settings drawer opens.)

The redesigned Reporting feature, under the newShareReportDrawer feature toggle, is enabled by default.

By configuring [time_picker] quick_ranges in your server config, you can set server-wide defaults specific to the data you are monitoring:

Copy

[time_picker]
quick_ranges = """[
{"from":"now-6s","to":"now","display":"Last 6 seconds"},
{"from":"now-10m","to":"now","display":"Last 10 minutes"},
{"from":"now-25h","to":"now","display":"Last 24 hours"},
{"from":"now/w","to":"now/w","display":"This week"},
{"from":"now-1w/w","to":"now-1w/w","display":"Last week"}
]"""

Thanks to Chris Hodges for contributing this feature!

We enhanced our custom currency format to solve this. Instead of the standard currency format, you can now use currency:financial:<unit>, which displays the complete numeric value instead of scaled abbreviations.

For example, currency:financial:$ will format large values like 1235667 as $1,235,667 instead of $1.24M.

Currency symbol positioning

The financial format also supports flexible symbol positioning:

Prefix (default): currency:financial:$ → $1,235,667

Suffix: currency:financial:€:suffix → 1,235,667€

The LogicMonitor Devices data source plugin enables you to query and visualize Device Instance Data, and to list Devices, Datasources, and Instances directly within Grafana. Available in public preview in Grafana Cloud and all supported versions of self-managed Grafana Enterprise.

You can now create alerts for each check in your Synthetic Monitoring application. For example, you can create an alert based on the number of check failures in a specific time window, with different settings for each one of your checks.

In addition to failed checks, you can also create an alert if a TLS certificate is set to expire within a certain number of days.

The default alert rules based on sensitivity thresholds can still be used in combination with per-check alert rules.

Per-check alert rules give users more granular control over their alerting needs, making it easier for users to configure the right alert threshold depending on the services they’re monitoring and the check settings. These alerts are powered by Grafana Managed Alerts.

To learn more, refer to the per-check alerts docs.

Here’s how it works:

• Restoring a pipeline version overwrites the configuration of the current pipeline.
• A restored pipeline is set to the same Active or Inactive state as the pipeline it overwrites.
• When a pipeline version is restored, its matching attributes can also be restored, or the matching attributes from the overwritten pipeline can be persisted in the restored version.

• Time period
• Component
• Cloud region
• Error level

Find any non-standard workloads using filters on the Workloads main page. Filtering includes:

• Argo Rollouts
• Strimzi Pod sets
• Unmanaged (or static) Pods
• CronJobs
• Bare Pods

In this case, you can see at a glance key data such as:

- How memory usage in the namespace compares to usage

- Which workloads could be contributing the most to the namespace’s memory usage

- Pods that could be incorrectly sized

This update introduces the ability to import rules directly from a Prometheus YAML file using the same alerting UI.

With early detection patterns, Adaptive Logs now shows log patterns without waiting to see how often they are queried. This feature allows you to proactively detect and filter out wasteful logs, ensuring only the most valuable telemetry is ever shipped to Grafana Cloud.

This change is part of our effort to provide:

• A simpler, more consistent experience across observability tools
• Faster access to relevant signals and workflows
• A foundation for deeper, context-aware troubleshooting powered by Grafana’s Knowledge Graph

We’re just getting started — more improvements will follow soon. If you have feedback, we’d love to hear it!

With SLO Reports, you can now:

• Consolidate multiple SLOs into a single, easy-to-digest report
• View individual and combined SLI (Service Level Indicator) metrics
• Understand weighted contributions of each SLO to the overall SLI performance

The combined SLI is calculated based on traffic-weighted contribution, meaning each SLO’s impact is proportionate to its share of total traffic for the reporting period. This gives teams a more accurate view of service health across distributed systems.

SLO Reports make it easier to monitor trends over time, share insights with stakeholders, and prioritize reliability efforts across high-impact services.

Benefits

• Incoming IRM calls and SMS messages appear with the contact name Grafana IRM instead of an unknown number.
• You can configure your phone to allow calls and messages from Grafana IRM to override DnD or silent mode.

How it works

1. In the IRM mobile app, go to Settings > Sync Grafana IRM contact.
2. Toggle the setting on to enable.
3. The app will create a contact named Grafana IRM using known IRM phone numbers.

The list of phone numbers is updated automatically whenever you open the app and new numbers are available. Once enabled, this setting adds a contact card with our verified SMS and call numbers.

Stay responsive, even when your phone is set to silent. Update your IRM mobile app today from the App Store or Google Play and try it out! Available in mobile app version 2.12.0 and later.

CPU tabs available at every level beyond the container give you:

• An edge for finding issues
• Capacity sizing insights to optimize performance

This feature lets you assign projects to a PLZ so that only tests inside of that project can be run on a specific PLZ.

For backward compatibility, any PLZ created before May 21st, 2025 is still able to run tests from any project.

PLZs created after that date have no projects added to it by default. To add a project to the PLZ, admins can edit the projects assigned to a PLZ by:

• Manually editing the Projects property of a PLZ via Testing & synthetics > Performance > Settings > Load zones.
• Using the Load zones REST API.

With RBAC, you can assign different roles and permissions to users and teams, ensuring the right people have the right access. This makes managing your monitoring workflows more secure and flexible.

How it Works

To assign roles and permissions to a user or team, simply follow the steps outlined in our guide for Assigning RBAC Roles. Here’s a quick example of how you can elevate a user’s permissions:

1. Assigning a Checks Writer Role:

From the Users and Access menu, we’re adding the Checks Writer role to the user checkswriter, who already has the Viewer role.

2. User Permissions in Action:

Once the role is assigned, the user will now have access to the Add New Check feature in the Synthetic Monitoring plugin.

3. Access Control in Action:

Other parts of the plugin, like the Probes creation flow, will remain read-only, respecting the user’s original Viewer role.

For detailed information on the available roles and more specific instructions, refer to the documentation.

TypeScript is a strongly typed language that extends JavaScript - all valid JavaScript code is also valid TypeScript. By adding static types, TypeScript helps you catch errors early in development, enabling you to identify bugs before your code runs.

To learn more about the different ways to author and run tests in the cloud, refer to our documentation.

• Custom incident statuses: Define additional phases, such as Investigating or Monitoring, to go beyond the default Active and Resolved statuses. Custom statuses help responders communicate progress more clearly and give stakeholders better visibility into the state of an incident.
• Custom metadata fields: Capture the information that matters most to your organization. Add custom fields, such as impacted services or involved teams, and even make them required to ensure consistent data collection across incidents.
• Private incidents: Limit visibility for sensitive incidents by declaring them privately. Private incidents automatically create a restricted Slack or Microsoft Teams channel and give you control over who has access throughout the incident.

To learn more, refer to the documentation on incident settings and private incidents.

ML Editors can:

• Create/edit/delete forecasts and outlier detectors
• Create/edit/delete holidays

ML Viewers can:

• View forecasts and outlier detectors

Sift Editors can:

• Start investigations
• Edit configurations

Sift Viewers can:

• View investigations

• Rank your metrics with new topk(n) and bottomk(n) functions to quickly get your highest and lowest ranking time series.
• Aggregate spans over time using sum_over_time() for built-in cumulative sums, such as total bytes, error counts.
• Fetch the latest traces first via the experimental most_recent=true query hint.

Ranking topk and bottomk functions for TraceQL metrics

When you’re looking at latency, error rates, or throughput across hundreds or thousands of services or endpoints, it’s easy to get lost in all the data. Previously, you’d have to pull back the full set of aggregates and then manually inspect or post‑process the results to find your worst offenders or best performers.

With topk(n) and bottomk(n), you can immediately narrow your focus to the top‑ or bottom‑ranked spans in a single, efficient query. This saves time and reduces the data volume you need to scan downstream.

Copy

{} | avg_over_time(span:duration) by (span:name) | topk(10)
{} | count_over_time() by (span:name) | bottomk(10)

These are second stage functions used, where:

• topk(n) returns the n series with the highest values from a first‑stage aggregation.
• bottomk(n) returns the n series with the lowest values.

For more details on these functions, refer to our documentation.

sum_over_time function for TracesQL metrics

With sum_over_time(), you can directly compute cumulative sums inside TraceQL, like total bytes transferred, total error counts, or resource consumption over time.

Copy

{}  | sum_over_time(span.http.response_content_length)

For each step interval, sum_over_time(attr) totals the values of attr across all matching spans.

For more details on this and other related functions, refer to our documentation.

Experimental query hint most_recent=true to retrieve the most recent traces

When troubleshooting a live incident or monitoring production health, you often need to see the absolute latest traces first. By default, Tempo’s query engine favors speed and returns the first N matching traces, which may not be the newest.

The most_recent hint ensures you’re always looking at the freshest data, so you can diagnose recent errors or performance regressions without missing anything due to early row‑limit cuts.

Copy

{} with (most_recent=true)

With most_recent=true, Tempo performs a deeper search across data shards, retains the newest candidates, and returns traces sorted by start time rather than stopping at the first limit hit.

For more details on this query hint, refer to our documentation.

You can define label-based segments via a GUI, API, or Terraform. Each segment will then receive its own recommendations; rules and exemptions can then be managed per segment.

To learn more, refer to the following video demo, as well as the Grafana Logs Drilldown documentation.

Logs Drilldown JSON viewer is available for all Grafana instances >= 11.6 and Loki instances >= 3.5.0

Connect your instance to a GitHub repository, and manage your dashboards as code directly from the Grafana UI. By enabling Git Sync, you will be able to manage dashboards the same way you manage your code. Dashboards will be versioned in Git, and edited through a PR workflow, so that it’s always possible to track changes.

This feature enables better control of your dashboards and enhance traceability of changes, without changing the way of editing a dashboard in the UI.

This is an experimental feature, so we don’t recommend using it in production environments.

Enable the provisioning and kubernetesDashboards feature toggles to try out Git Sync for Grafana Dashboards today!

Using these tools and APIs, you can import data source-managed alert rules from Prometheus-compatible systems (Prometheus, Mimir, Loki) into Grafana as Grafana-managed alert rules. This simplifies migrating from data source alerts to Grafana Managed Alerts while preserving behavior of the rules. Alert rules can be imported from an easy-to-use UI tool or through API endpoints compatible with Mimirtool.

View the documentation to find out more about how the tool works and how you can configure your import options.

Read the documentation for more information about how to set a “Keep firing for” duration so you can use the Recovering state to eliminate unwanted noise.

Enhancements:

Performance: On a simulated large table of 41,400 rows and 17 columns, we see the following improvements:

Cell text wrap: The new table visualization implements a dramatically faster text wrap for larger tables and allows toggling text wrap on a field-by-field basis using overrides. We’re using a new open source text wrap library built by our own dataviz grafanista (uWrap)! We have removed the text wrap feature from its beta state as a result.

This new version of the table visualization is in public preview. To try the new table, turn on the tableNextGen feature toggle. If you use Grafana Cloud, just ask Grafana Support for the new table.

The issue only exists for self-managed users who want to roll back to a previous version of Grafana and users who provision dashboards as code. Grafana v12 has a mechanism to resolve both versions. 

Note that if you’ve requested early access to dynamic dashboards through this form, this will enable an experimental feature and should be used according to our guidelines.

• resource_cache_size
• query_cache_size

While available metrics are not officially documented, you may still use them to get information on the state of your instance.

Migration/mitigation

You’ll continue getting data by using the cache_size metric, but we recommend you update your dashboards to use the resource_cache_size and query_cache_size metrics instead.

This affects plugin developers that are using UI extensions and haven’t migrated to the new APIs yet. Plugins trying to use the deprecated APIs in Grafana v12 throw an error.

Migration/mitigation

For details on how to migrate to the new APIs, refer to the migration guide.

In Grafana v11.2, we added a new failWrongDSUID feature toggle that is turned off by default. When enabled, the REST APIs and provisioning reject any requests to create or update data source instances that have an incorrect UID.

We are enabling this toggle by default in Grafana v12.

Correct UIDs may only contain latin characters, numbers, and dashes. Find the exact Regex definition in our codebase here.

Migration/mitigation

You can turn off the failWrongDSUID feature toggle, but keep in mind that this behavior will become standard and not configurable within the next 6 months.

The Grafana version check is re-enabled in Grafana v12 when installing plugins.

Grafana Cloud users are not affected as they can’t use the CLI to install plugins. We expect there are only a handful of cases for self-managed instances.

What to do if I want to ignore the compatibility check?

We do not recommend installing plugins declared as incompatible. But, if you really have to force install a plugin despite it being declared as incompatible, refer to the Installing a plugin from a ZIP guidance.

You can filter, do math, and join across different data source types at run time.

For example:

• Compare Big Table business data against operational metrics in Prometheus.
• Use the GitHub data source and refine issues or PR display.
• Turn the above into alerts, for example issues that haven’t been handled, or dynamic alerts thresholds based on ARR.

Migration/mitigation

RED metrics are supported via Traces Drilldown and can be viewed with more advanced context than via the older metrics summary API. For all users, the aggregate by section will no longer be present in the Tempo data source search tab. Additionally, any user that had been using the aggregate by functionality in the Tempo data source in a dashboard can manually change the Tempo query.

For more information, refer to our blog post and documentation.

If any, the impact would be that actions are not displayed in the context menu of the table visualization.

Migration/mitigation

Actions in the table visualization are supported using the Actions cell type. There is no migration in place, as we risk losing the cell content if we migrate to an Actions cell. If you consider that approach appropriate, however, you can manually change the cell type and gain access to actions, or add another column that supports actions.

We are removing this experimental page from Grafana Cloud since we are replacing our current feature toggle system with one that is more robust, and the new system is incompatible with the UI.

This change affects administrators who use this UI to view and manage toggles. However, the removal of the UI won’t affect which toggles are enabled or disabled on your stack.

Migration/mitigation

You don’t need to take any action. If you want a list of your active feature toggles at any time, open a support ticket.

Previously, provisioned users needed to individually sign in to Grafana to exist within the platform, and administrators faced the time-consuming task of creating and managing teams - either manually or via provisioning. This process was complicated and often presented a bottleneck for new team members getting up and running quickly.

This release introduces SCIM (System for Cross-domain Identity Management) integration, enabling seamless synchronization of Grafana Teams directly from your Identity Provider (IdP).

Key Benefits:

• Instant Team Onboarding: Onboard entire teams to Grafana instantly, ensuring they are ready to be added to on-call schedules and receive critical push notifications from day one.
• Simplified Administration: Eliminate the need for manual team creation and management within Grafana. Leverage your existing IdP to maintain team structures.
• Broad Compatibility: SCIM works with Okta SAML and Azure AD SAML.
• Automated Access Management: Ensure timely revocation of access when team members leave your organization, directly managed through your IdP.

Enable the enableSCIM feature toggle to start using SCIM to simplify team management and accelerate the onboarding process for users!

The new APIs make Git Sync possible, and are the key change that power our new Terraform provider and the creation of the GrafanaCTL CLI tool. Right now, only dashboards, folder, and other few elements like announcement banners and playlists are leveraging this new model, and we plan to expand it to all Grafana resources.

For more information, you can read our documentation.

Simplifying Your Migration Journey

Migrating from OSS or Enterprise Grafana to Grafana Cloud has traditionally been complex, requiring technical knowledge of Grafana’s HTTP API and time-consuming manual processes. The Grafana Cloud Migration Assistant eliminates these challenges by providing a user-friendly interface that automates the migration of your resources. No coding required — it securely handles the transfer in just a few easy steps.

What’s Included

The Migration Assistant provides:

• An intuitive UI with real-time updates on your migration status
• Secure, automated transfer of dashboards, folders, alert rules, plugins, and more
• Seamless migration of data source and plugin credentials

With the general availability release, you can now confidently migrate your OSS instance or production environments with enterprise-grade reliability and performance.

Enhanced Capabilities

Since our public preview, we’ve refined the Migration Assistant based on valuable user feedback. You can now select which resources to migrate, and perform sorting and filtering on the migration results table. We also offer improved stability, faster migration speeds, and improved error handling to make your transition to Grafana Cloud faster, more efficient, and less error-prone.

Get Started Today

Ready to make the move? Explore our migration guide to learn more about the Cloud Migration Assistant today and begin your effortless transition to Grafana Cloud.

For questions or assistance, contact our support team or join the conversation in our community forums.

Centralized Signal Analysis for Efficient Investigations

• Add Panels from Any Signal: Investigations breaks down the silos between our Drilldown applications. You can now seamlessly add panels from Metrics, Logs, and Traces Drilldown directly into a single investigation view.
• Side-by-Side Comparison Across Time: Gain deeper insights by comparing the behavior of your signals across two distinct timeframes, all within the same view. This allows for the rapid identification of correlations and anomalies.
• Effortless Collaboration: Investigations make it easy to share your findings and the context of your analysis with other team members involved in the same task, fostering better collaboration and faster resolution.

Investigations provides a central hub to correlate and analyze diverse signals across the otherwise siloed Drilldown applications, leading to more efficient and effective troubleshooting.

Enhanced Visualization and Filtering Capabilities

• Multiple Include Filters: You can specify multiple inclusion filters within your Drilldown queries, allowing for more nuanced and targeted log retrieval.
• Regex Support: We’ve added support for regular expressions in label and field queries, providing robust and flexible pattern matching for in-depth log analysis.
• JSON Visualization: Explore and understand your JSON-formatted log entries with our new dedicated JSON visualization, making complex data structures easier to read and interpret.
• Service Selection Pagination: The service selection is now paginated, making it easier for users to view and discover service related data.
• Sort order changes query direction: Sort order and query direction are now linked, using a new query direction “Forward” from Loki. Setting “Oldest first” will query your log data in the order of oldest log to most recent. Setting “Newest first” will query your logs from a “Backwards” direction (i.e. newest log line to oldest).

With these new features in Logs Drilldown, we’re empowering you to gain clearer insights from your log data by supporting a wider range of formats and offering more powerful filtering and visualization options.

Faster Filtering

• Prefix and Suffix Filters: Refine your metric exploration with the new prefix and suffix filter options, enabling more precise targeting of the data you’re looking for.
• Group By Labels: Organize and understand your metrics more effectively by grouping them based on their labels. This provides valuable context and simplifies analysis.
• Sort Options: Tailor your view by sorting metrics based on recent history, associated dashboards, and related alerts. This allows you to prioritize and investigate efficiently.

Improved UI

• Collapsible Sidebar: We’ve introduced a collapsible sidebar, allowing you to focus on the data that matters most by providing more screen real estate for filtering.
• Better Metric Count: We now provide a real-time count of the metrics displayed after each filtering action, giving you immediate feedback on the impact of your selections.

With these enhancements to Metrics Drilldown, we continue our commitment to providing powerful tools to streamline your observability workflows and gain deeper insights into your systems.

The editors_can_admin setting in Grafana allows users with the Editor role to create and manage Grafana Teams. Editors in instances with this configuration setting no longer have these extra privileges, but they are still able to manage teams they already created while the setting was active.

We’re removing the settings because of the impact they would have on our new access engine. We’re not able to support these settings with the engine. We’re also considering these settings to be outside of the current engine, making it unsafe to be give continued support to these settings.

Migration/mitigation

There are two alternatives for editors_can_admin:

• Update Grafana Editors to Grafana Admins.
• Grant team creation permissions to users who need to manage teams. Note this is not possible for OSS users.

• Logs filtered by the same trace or service fields
• Metrics focused on the exact service, endpoint, or error condition
• Profiles for CPU, memory, or heap snapshots tied to a span
• Dashboards, runbooks, tickets, or any external URL that enriches your workflow

All it takes is a quick rule in Correlations. Simply define which span or trace fields to use as variables, point to your target, and watch contextual links appear in the Explore trace view.

Available today for all Grafana Cloud stacks running Grafana Cloud Traces.

Native histograms are a metric type in Prometheus that can produce, store, and query high-resolution histograms of observations. They usually offer higher resolution and more straightforward instrumentation than classic histograms.

Since the private preview, we have progressed on a new pricing model and added the out-of-order feature. These newly added supports add more value and accuracy to this feature. Please give it a try and let us know how you like it!

k6 browser checks allow you to go beyond traditional monitoring, enabling you to test and validate real user flows and gain insight into performance in ways you can’t with HTTP-level checks. With browser checks, you’re not just monitoring performance at a protocol level; you’re simulating user interactions that provide you with a much richer view of your app’s behavior in production.

With k6 browser checks, you can control a headless Chrome browser using a Grafana k6 script. Powered by the k6 browser module, these checks can:

• Simulate critical user journeys from across the globe
• Measure real-world frontend performance with Web Vitals
• Catch regressions before they reach production
• Use the same scripts across Grafana Cloud k6 and Synthetic Monitoring

To learn more, refer to the Create a k6 browser check and k6 browser check docs.

• Attribute costs
• Track ownership
• Apply consistent metadata

Especially valuable for teams without direct access to AWS tagging, this feature:

• Centralizes label management
• Reduces duplication
• Removes the need for manual, error-prone configuration

With this latest update, you can now dive into the full incident timeline directly from your phone. Instantly catch up on everything that’s happened: comments, notes, attachments, and key activity — all organised chronologically.

When you open an incident in the IRM Mobile app, you’ll now see three tabs:

• Overview: basic incident info like roles, participants, and creation details
• Timeline: a complete log of all activity, decisions, and updates throughout the incident
• Attachments: a simple list of all the files and items attached to the incident

These new tabs make it faster and easier to stay informed and take action, even while you’re on the move.

We’re continuing to evolve the IRM Mobile app based on your feedback — and we’re just getting started. Update your app today from the App Store or Google Play and check it out!

Key highlights

• Queryless exploration made simple: Traces Drilldown eliminates the need to write complex queries. Users can instantly access granular trace-level insights directly from their dashboards. This “queryless” design allows you to effortlessly connect the RED (Rate, Errors, Duration) golden signals with precise trace details.
• Seamless navigation: Enjoy a smooth transition between the high-level trace overview and deep dive span views. This unified navigation experience ensures that you maintain context as you shift from summary views to detailed analysis. This advanced drilldown capability improves incident analysis by providing the contextual insight needed for faster root cause identification.
• Built-in investigative tools: The Traces Drilldown interface features integrated filtering, highlighting, breakdowns, comparisons, and root cause correlation. These built-in tools enable rapid trace assessment by quickly spotlighting abnormalities without the need for external analysis tools.
• Detailed trace visualization: Once you’ve narrowed down your traces of interest, the trace view high-fidelity spans timeline reveals key metadata, duration, and status indicators, helping you quickly identify the origin of performance bottlenecks and errors.
• Unified observability experience: Traces Drilldown connects traces with logs, metrics, and profiles for a cohesive observability workflow. Effortlessly transition from an anomaly in logs or metrics into detailed trace views, enabling you to inspect spans and duration breakdowns to swiftly uncover underlying issues.

What’s new for GA

Exemplars

Now, when inspecting your traces, you can take advantage of integrated exemplars embedded directly within the trace metrics. Exemplars highlight representative spans that capture critical events or anomalies, providing a quick visual summary of key operations within your trace data. They help you correlate and prioritize areas that warrant further investigation, all without having to manually sift through extensive trace details.

Streaming

With support for TraceQL query streaming, Traces Drilldown now delivers partial results as they come in, so you no longer have to wait for all results to finish processing before starting your analysis. This capability enables you to review metrics and traces immediately, speeding up root cause analysis and helping you resolve issues more quickly. By receiving incremental results, you can begin exploring critical trace details as soon as they’re available, ensuring a more responsive and dynamic observability experience.

Bug fixes, fit and finish, and quality of life improvements

This release also brings a host of stability and usability refinements. We’ve addressed a range of bugs reported during the public preview phase and made targeted adjustments to the user interface, ensuring smoother navigation and a more polished overall experience. These quality of life enhancements, spanning from improved interactive elements to performance optimizations, empower you to conduct trace analysis with greater efficiency and confidence.

What this means for you

• Whether you’re tracing a request through your microservices or diagnosing performance anomalies, the Traces Drilldown experience streamlines investigations by simplifying your workflow. Start from one of the golden monitoring signals, pinpoint an issue, and drill down into traces through grouping and comparing, all without having to write custom queries.
• The enhanced, queryless interface accelerates your time-to-insight by reducing friction for operators and developers alike, enabling faster resolution of incidents and more effective optimization of distributed architectures.
• As part of the broader evolution of Grafana’s queryless UI, Traces Drilldown unifies your observability by working seamlessly with the improvements across logs, metrics, and profiles. This creates a truly integrated observability suite that meets users where they are.

With Traces Drilldown now GA, we continue our commitment to empowering modern observability teams, simplifying your investigative process and ensuring you have the right tools to maintain high performance across complex distributed systems.

Categorize your collectors with meaningful attributes, such as environment type, department, or location. Then inject these attributes into configuration pipelines using the syntax argument.attributes.value["ATTRIBUTE_KEY"]. With attribute injection, you can

• Create a single pipeline that works across different environments.
• Dynamically set labels based on collector attributes.
• Reduce the number of pipelines needed to manage your fleet.
• Make configurations more maintainable and scalable.

• Renamed the Organization settings to Admin & API.
• Added Admin-level setting to enable or disable the availability of notification channels (such as Slack or MS Teams) across all user profiles.
• Deprecated OnCall API tokens in favor of Grafana Cloud service accounts. Existing OnCall API tokens can still be used, viewed, and deleted, but new tokens can no longer be generated.

When a team is assigned to a service and has an active on-call schedule, the service details page shows:

• Who is currently on call
• Their local time and whether they’re within working hours
• Quick-access actions like Slack DM and escalation

From the same view, you can:

• Escalate to a team or individual
• Declare an incident
• Send a Slack DM to the on-call responder

Escalations and incidents automatically include the service_name label and a link back to the service details page for added context.

With this new architecture based on Scenes, we’re simplifying the schema to make it easier to understand and support powerful new features, like dynamic dashboards, tabs, and canvas-style layouts. The new schema also aims to simplify dashboard management, supporting the as-code approach of Git provisioning.

In the UI, the new schema is automatically enabled with Dynamic dashboards feature and also requires the new dashboards API feature toggle, kubernetesDashboards, to be enabled.

Note that this is an experimental feature and should not be used in production environments, as it might result in an irreversible loss of data.

What does this mean for new dashboards?

With the new schema enabled, you can build dynamic dashboards that support tabs, conditional logic, flexible panel layouts, and much more.

What does this mean for my old dashboards?

You can still open and view your pre-existing dashboards. Upon saving, they’ll be updated to the new schema, where you can take advantage of the new features and functionalities.

What if I disable the dynamic dashboards feature toggle?

Depending on how the dashboard was built, it will behave differently: 

• Dashboards built on the new schema through the UI - View only
• Dashboards built on Schema v1 - View and edit
• Dashboards built on the new schema with Terraform or the CLI - View and edit 

What if I disable both the dynamic dashboards and dashboards API feature toggles?

You’ll be unable to view or edit dashboards created or updated in the new schema.

Import/export

From the UI, dashboards created on the new schema can be exported and imported as other dashboards. When you export them to use in another instance, references of data sources are not persisted, but data source types are. You’ll have the option to select the data source of your choice in the import UI.

Now with Pause Adaptive Logs, end users of logs can temporarily ingest the logs they need for the service, application, or cluster they care about, without needing to configure anything in Adaptive Logs or work with their centralized team. Get all the cost savings benefits of Adaptive Logs, while making sure your dev teams aren’t stranded during an Incident!

Pause Adaptive Logs is in public preview for Grafana Cloud customers, and will be available in all Cloud stacks by April 21. For more information, see our documentation.

You can now manage incident roles directly from your phone!

Quickly update who’s doing what during an active incident - Whether you need to assign a critical role quickly or adjust assignments as the situation evolves, you can do it all within the IRM mobile app.

Here’s an overview of what you can do:

• Quickly assign & reassign roles by searching and selecting users
• Easily unassign users from roles when they’re no longer needed
• Jump straight into Slack to message an assigned user

Update your IRM mobile app today from the App Store or Google Play and experience the improved incident role management!

This opens the Change theme drawer, where you can view all themes:

Now you can see your selected theme. Enjoy!

This feature is automatically enabled in Grafana Cloud. Enable the `grafanaconThemes` feature toggle in Grafana to use this feature. You can find updated documentation in User preferences and in Organization preferences. Let us know what you think in our feedback survey.

More efficient navigation and consumption

• Tabs for better organization: Build well-structured dashboards with tabbed layouts for different contexts and user groups.
• Conditional rendering: Show or hide visualizations based on variable selections or data availability, ensuring relevance. 
• Dashboard content outline: Navigate dashboards fast with this tree structure, making it easier to jump between sections.

Powerful dashboard editing

• New flexible panel layout options: Use custom or auto-grid panel layouts with adjustable settings, like column count, for adaptive dashboard behavior.
• Tabs and row grouping: Group panels into rows or tabs based on your needs. Nested grouping to support complex visual needs is also supported.
• Context-aware editing: Quickly modify tabs or rows, panels, and variables without diving into full edit mode.
• Intelligent side pane: The side pane user interface adapts to your selection, enabling fast actions like duplicating elements or setting layout options.

This is an experimental feature and should not be used in production environments, as it might result in a loss of data.

We’re excited to introduce a new controls component for Logs in Explore, designed to make better use of the layout around logs, and give users greater flexibility when working with log data.

This update allows users to configure key display settings for a more tailored viewing experience, such as:

• Log direction
• Deduplication strategy
• Timestamp visibility
• Line wrapping
• JSON parsing

Users can also access the download button for easy export of log data.

Lastly, a new log level filter makes it simpler than ever to narrow results by severity, helping users focus on what matters most.

These enhancements make log exploration faster, more efficient, and more customizable.

ML Editors can:

• Create/edit/delete forecasts and outlier detectors
• Create/edit/delete holidays

ML Viewers can:

• View forecasts and outlier detectors

Sift Editors can:

• Start investigations
• Edit configurations

Sift Viewers can:

• View investigations

Now, when executing k6 tests in the cloud from a GitHub Actions workflow, the pull request comments will automatically display a rich test summary. This includes key performance metrics like response times, request rates, pass/fail statuses and comparison with the baseline test runs—providing instant, actionable insights without leaving your development workflow.

Stay on top of your application’s performance with improved visibility and efficiency!

You can view all changes across all pipelines, review changes to a specific pipeline, compare differences between change events for a pipeline, or copy configuration syntax to revert to an earlier version of a pipeline.

RBAC provides a standardized way of granting, changing, and revoking access when it comes to viewing and modifying Grafana resources, such as dashboards, reports, and administrative settings.

Please note that Frontend Observability RBAC only includes a set of basic roles and not fine-grained access.

Frontend applications often serve a globally distributed user base, meaning that your users can be spread across multiple geographical regions and access your application from diverse locations. In some instances, a user’s location can affect the performance of web applications.

With this new feature, Frontend observability can help you determine the approximate location of your users and understand if your application experiences a performance degradation due to user location in relation to the location where the application is deployed. It also can help you visualize demographic information, such as the region that appears most prominent across a user base.

You can achieve these insights without compromising data security or compliance with privacy laws like GDPR, CCPA, and ePrivacy directive. With Geolocation Insights, you can filter, segment, and analyze frontend telemetry based on location data to optimize web performance, localize content, and tailor experiences, all while incorporating privacy by design.

k6 extensions are libraries that expand the k6 core functionality. For example, the faker k6 extension enables users to generate random fake data, which is not available on k6 by default. Extensions allow the k6 open source project to remain small and focused on its core functionality, but still provide users with a way to add new features to cover any use case, such as supporting a new network protocol.

You can now seamlessly import and use a pre-approved set of k6 extensions in your Grafana Cloud k6 tests. Here is an example script that uses the k6/x/faker extension:

js Copy

import faker from "k6/x/faker";

export const options = {
  cloud: {
    name: 'Extensions in Grafana Cloud k6',
  },
}

export default function () {
  console.log(faker.person.firstName());
}

We also display the versions of k6 and extensions that your test runs are using, with links that take you to the source code, to help you debug issues faster.

Refer to the docs to learn about how to leverage this feature and what extensions are initially available in the cloud.

You can now create alerts for each check in your Synthetic Monitoring application. For example, you can create an alert based on the number of check failures in a specific time window, with different settings for each one of your checks.

In addition to failed checks, you can also create an alert if a TLS certificate is set to expire within a certain number of days.

The default alert rules based on sensitivity thresholds can still be used in combination with per-check alert rules. Per-check alert rules give users more granular control over their alerting needs, and make it easier for users to configure the right alert threshold depending on the services they’re monitoring and the check settings.

To learn more, refer to the per-check alerts docs.

No more rushing back to your laptop when you notice a typo in an incident title or need to update the severity level while you’re away from your desk. Simply open the Grafana IRM mobile app, navigate to the incident, and make your changes on the spot.

You can now:

• Fix that embarrassing typo in your incident title or description
• Update the status as the situation evolves
• Change severity when you realize it’s worse (or better) than you thought
• Add or remove labels from the incident

The update is available now - just grab the latest version from the App Store or Google Play.

Questions or feedback? Let us know what you think!

• AI-powered flame graph interpretation
• Incident auto-summary
• Dashboard panel title and description generation
• Explanations of error log lines in Sift

We now support the usage of Anthropic’s API alongside already existing support for OpenAI’s API, OpenAI-compatible APIs and custom APIs.

Having flame graphs explained: one of the capabilities in the LLM plugin

Previously, adding gzip manually as a request header caused parsing issues. With this update, requests automatically include the Accept-Encoding: gzip header, ensuring smoother data retrieval and visualization.

Admins can configure these settings at the data source level, ensuring that metadata is passed consistently while preventing users from overriding values in individual queries. Since different APIs require metadata in different formats, this feature offers flexibility in how values are forwarded.

To maintain security and compatibility, metadata is only handled at the backend and is not automatically forwarded in headers. Additionally, this does not rely on Grafana’s global send_user_header setting, allowing per-instance configuration.

Previously, Infinity only supported GET and POST, as it was designed for read-only queries. However, some use cases require modifying data via PATCH, PUT, or DELETE. Since these methods can perform destructive actions, they are disabled by default.

To enable this, toggle allowDangerousHTTPMethods in your data source configuration.

📩 Automated email reminders – Receive email notifications before a token expires, ensuring a seamless transition.

🛠 Improved token expiry visibility – Easily navigate expiring and expired tokens with clear token warning labels, and a red badge alert in the Cloud Access Policy user interface.

With Debug Metrics, you have:

• Faster issue detection when data looks off
• Immediate insights to restore accuracy
• More confidence in your data-driven decisions

Not only can you troubleshoot what data or labels are missing, you can easily understand and learn what data is driving any panel.

Previously, AWS PrivateLink only supported connectivity to VPC endpoint services in the same region. Connecting to services in a different region required setting VPC Peering between both regions, and this was complicated in some environments.

This launch allows you to connect directly to Grafana Cloud endpoint services hosted in other AWS Regions over PrivateLink endpoints.

• One app for on-call and incident response – On-call and incident settings are now managed in a single location. No more switching between separate apps in Grafana Cloud.
• New My IRM dashboard – Get a complete view of ongoing alerts and incidents in one place.
• Seamless transition – Existing OnCall and Incident users retain all data; all on-call schedules, integrations, and settings remain intact.

The Grafana Cloud IRM app is gradually rolling out between March 17-24, 2025, and will be available in all Cloud instances by the end of March.

Additional buttons are available on any Cluster detail page for you to immediately access the namespaces and workloads on the Cluster.

• Provision your collectors and configuration pipelines with Terraform.
• Deploy and configure collectors as native Kubernetes resources.
• Add pre-built configuration pipelines from the integration catalog.
• Learn how to streamline your remote configuration workflows to help you save time and money.

Get started with Grafana Fleet Management today!

We’re pleased to announce an improvement to the Grafana plugin catalog that benefits both Grafana users and plugin developers. By introducing standardized links on plugin details pages, we’re making it easier for users to engage with developers, and find the essential information they need to get the most out of a plugin. Developers, in turn, will gain valuable feedback and support from their audience.

Previously, plugin details pages in the Grafana Catalog relied on custom links provided by developers, leading to inconsistencies in information being available. These are still configurable, but are now augmented with standard links for key elements such as raising an issue, or navigating to the source code repository if the plugin is open source.

Action required for plugin developers

We recommend plugin developers check the inferred links by viewing their plugin in the catalog. You can update these by providing values in your plugin.json metadata file and submitting a new update to the catalog. Consider removing redundant custom links which are now served by the standardized links.

Next steps

We are continuing to evolve the plugin catalog experience to make it easier to understand a plugin’s capabilities and support their developers. In a future update we are looking to provide plugin developers the means to advertise github sponsorship or other options so that users can help support the continued enhancement and maintenance of community plugins they depend on.

We welcome any and all feedback or suggestions for the plugin catalog, you can let us know your thoughts by raising an issue in the main Grafana repository.

The following examples show a comparison of before and after switching to WebGL.

Before

After

Note: You might notice a slight change in styling due to the change in rendering technology. As always, we look forward to your feedback!

Alerting has added a new feature that allows you to review, compare, and restore previous alert rules.

• You now can restore old rule versions from the versions tab in the detail view.

• For Immutable rules (rules that are provisioned, federated, or rules that come from a plugin), versions cannot be restored, only compared.

• If when trying to restore a version, is not possible due to different changes in the environment, the UI redirects the user to the alert rule form with the version to recover. This allows you to review, correct the data, and save the rule.

LBAC (Label Based Access Control) for data sources - metrics is our second step towards seamless management of access for Mimir metrics. Each team views the same data source filtered by their team’s label permissions.

Feature highlights

• Teams can view queries to the same data source with different LBAC rules applied
• Configurable using API and the UI
• Simplified LBAC notation is automatically converted into correct custom headers

Best practices

• We recommend you only add query permissions for teams that will have LBAC rules and remove default Viewer and Editor query permissions.
• As an initial setup, we recommend defining as few rules as possible for each team, making sure that they’re additive rather than negating one another.
• For validating rules, we recommend testing each rule in Explore view. This allows you to see the logs that would be returned for the specific rule.

LBAC (Label Based Access Control) for data sources - logs (previously called Team LBAC) is our first step towards seamless management of access for Loki logs. Each team views the same data source filtered by their team’s label permissions.

Feature highlights

• Teams can view queries to the same data source with different LBAC rules applied
• Configurable using the API and the UI
• Simplified LBAC notation is automatically converted into correct custom headers

Best practices

• We recommend you only add query permissions for teams that will have LBAC rules and remove default Viewer and Editor query permissions.
• As an initial setup, we recommend defining as few rules as possible for each team, making sure that they’re additive rather than negating one another.
• For validating rules, we recommend testing each rule in Loki Explore view. This allows you to see the logs that would be returned for the specific rule.

• Time series
• Bar chart
• Candlestick
• State timeline
• Status history
• Table
• Trend
• XY chart

You can find actions under the Data links and actions section of the panel editor pane (previously the Data links section).

When you use dashboard variables in transformations, the variables are automatically interpolated before the transformations are applied to the data.

For an example, refer to the Filter fields by name transformation documentation.

As part of this update, the transformationsVariableSupport feature toggle has been deprecated.

To learn more about dashboard variables in transformations, refer to the following video demo:

• Time series
• Bar chart
• Candlestick
• State timeline
• Status history
• Trend
• XY chart

You can only have one One click data link, so if you enable it for one, the feature is automatically disabled for others.

Choose from a list of platform-specific integration templates to automatically build configuration pipelines in Fleet Management and then assign attributes to match those pipelines to specific collectors.

With integration templates in Fleet Management, you can remotely configure all your collectors in just a few clicks.

Use Terraform to preregister new collectors for the Fleet Management service, add remote attributes to collectors that are already registered, and build configuration pipelines to remotely configure your fleet. With Terraform and Fleet Management, you can create a stable, consistent, scalable observability setup.

With the promotion to general availability, the One click switch has been moved into the dialog box where you configure the data link or action, at element level. You can only have one One click data link or action, so if you enable it for one, the feature is automatically disabled for others.

• Prevent slowdowns: Get alerts for high CPU usage in AWS EC2, Azure VMs, and GCP Compute Engine.
• Keep databases running smoothly: Avoid failures with alerts for deadlocks, failed connections, and disk space limits.
• Optimize storage and messaging: Stay ahead of storage constraints and ensure messages are delivered on time.

Whether you’re running on AWS, Azure, or GCP, you can fix issues faster because of real-time notifications, including:

• High CPU usage
• Storage limits
• Failed connections

Start monitoring smarter today with Cloud Provider Observability in Grafana Cloud!

Why Service Accounts:

• Role-Based Access Control (RBAC) – Assign roles (for example, Viewer, Editor, Admin) and scope access to specific teams for improved security.
• Audited Automation – Service Accounts are integrated with Grafana’s identity management system, ensuring all actions are logged for better tracking and compliance.
• Seamless Token Rotation – Manage multiple tokens per service account effortlessly, allowing for smooth rotation without downtime.

API keys are no longer supported. To learn more about this change and how it enhances security, check out our blog post.